1 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
5 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
7 * tree-if-conv.c (is_cond_scalar_reduction): New function.
8 (convert_scalar_cond_reduction): Likewise.
9 (predicate_scalar_phi): Add recognition and transformation
10 of simple conditioanl reduction to be vectorizable.
12 2014-05-01 Marek Polacek <polacek@redhat.com>
15 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
17 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
19 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
20 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
21 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
22 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
23 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
24 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
25 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
26 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
28 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
30 * config/arc/arc.opt (mlra): Move comment above option name
31 to avoid mis-parsing as language options.
33 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
35 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
36 * config/sol2.h: ... here.
37 * config/sol2-10.h: Remove.
39 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
40 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
41 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
42 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
43 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
44 * config/sol2.h: ... here.
45 (SECTION_NAME_FORMAT): Don't redefine.
46 (STARTFILE_ARCH32_SPEC): Rename to ...
47 (STARTFILE_ARCH_SPEC): ... this.
48 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
49 * config/sparc/sol2.h: ... here.
50 (SECTION_NAME_FORMAT): Don't undef.
51 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
52 (SUBTARGET_EXTRA_SPECS): Remove.
53 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
55 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
56 (MD_STARTFILE_PREFIX): Remove.
57 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
58 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
59 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
60 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
61 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
62 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
63 * config/i386/sol2.h: ... here.
64 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
65 * config/i386/sol2-bi.h: Remove.
66 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
67 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
69 * config/i386/t-sol2-64: Rename to ...
70 * config/i386/t-sol2: ... this.
71 * config/sparc/t-sol2-64: Rename to ...
72 * config/sparc/t-sol2: ... this.
74 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
75 sol2_tm_file_head, sol2_tm_file_tail.
76 Include ${cpu_type}/sol2.h before sol2.h.
78 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
79 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
80 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
81 Reflect i386/t-sol2-64 renaming.
82 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
83 Reflect sparc/t-sol2-64 renaming.
85 2014-04-30 Richard Biener <rguenther@suse.de>
87 * passes.c (execute_function_todo): Move TODO_verify_stmts
88 and TODO_verify_ssa under the TODO_verify_il umbrella.
89 * tree-ssa.h (verify_ssa): Adjust prototype.
90 * tree-ssa.c (verify_ssa): Add parameter to tell whether
91 we should verify SSA operands.
92 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
93 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
94 whether we should verify whether not throwing stmts have EH info.
95 * graphite-scop-detection.c (create_sese_edges): Adjust.
96 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
97 * tree-eh.c (lower_try_finally_switch): Do not add the
98 default case label twice.
100 2014-04-30 Marek Polacek <polacek@redhat.com>
102 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
103 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
104 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
105 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
107 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
109 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
110 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
111 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
112 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
113 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
114 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
115 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
116 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
118 2014-04-29 David Malcolm <dmalcolm@redhat.com>
120 * tree-cfg.c (dump_function_to_file): Dump the return type of
121 functions, in a line to itself before the function body, mimicking
122 the layout of a C function.
124 2014-04-29 Jakub Jelinek <jakub@redhat.com>
126 PR tree-optimization/60971
127 * tree-tailcall.c (process_assignment): Reject conversions which
130 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
132 * calls.c (initialize_argument_information): Always treat
133 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
134 (expand_call): Likewise.
135 (emit_library_call_calue_1): Likewise.
136 * expr.c (PUSH_ARGS_REVERSED): Do not define.
137 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
140 2014-04-29 Nick Clifton <nickc@redhat.com>
142 * config/msp430/msp430.md (umulsidi): Fix typo.
143 (mulhisi3): Enable even inside interrupt handlers.
144 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
145 bigger return address pushed in large mode.
147 2014-04-29 Nick Clifton <nickc@redhat.com>
149 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
150 (arc_init_reg_tables): Use a machine_mode enum to iterate over
152 * config/m32r/m32r.c (init_reg_tables): Likewise.
153 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
154 enum to hold the modes.
156 2014-04-29 Richard Biener <rguenther@suse.de>
158 * dominance.c (free_dominance_info): Add overload with
160 (dom_info_state): Likewise.
161 (dom_info_available_p): Likewise.
162 * basic-block.h (free_dominance_info, dom_info_state,
163 dom_info_available_p): Declare overloads.
164 * passes.c (execute_function_todo): Verify that verifiers
165 don't change dominator info state. Drop dominator info
166 for IPA pass invocations.
167 * cgraph.c (release_function_body): Restore asserts that
168 dominator information is released.
170 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
172 * doc/invoke.texi: Fix typo.
173 * tree-vrp.c: Fix typos.
174 * gimple.c (infer_nonnull_range): Reorder operands of an &&
177 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
179 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
181 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
183 * config/aarch64/aarch64-builtins.c
184 (aarch64_types_storestruct_lane_qualifiers): New.
185 (TYPES_STORESTRUCT_LANE): Likewise.
186 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
187 (st3_lane): Likewise.
188 (st4_lane): Likewise.
189 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
190 (vec_store_lanesci_lane<mode>): Likewise.
191 (vec_store_lanesxi_lane<mode>): Likewise.
192 (aarch64_st2_lane<VQ:mode>): Likewise.
193 (aarch64_st3_lane<VQ:mode>): Likewise.
194 (aarch64_st4_lane<VQ:mode>): Likewise.
195 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
196 * config/aarch64/arm_neon.h
197 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
198 use new macro arguments.
199 (__ST3_LANE_FUNC): Likewise.
200 (__ST4_LANE_FUNC): Likewise.
201 * config/aarch64/iterators.md (V_TWO_ELEM): New.
202 (V_THREE_ELEM): Likewise.
203 (V_FOUR_ELEM): Likewise.
205 2014-04-28 David Malcolm <dmalcolm@redhat.com>
207 * doc/gimple.texi: Replace the description of the now-defunct
208 union gimple_statement_d with a diagram showing the
209 gimple_statement_base class hierarchy and its relationships to
210 the GSS_ and GIMPLE_ enums.
212 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
214 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
215 * config/aarch64/aarch64.c
216 (aarch64_cannot_change_mode_class): Weaken conditions.
217 (aarch64_modes_tieable_p): New.
218 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
220 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
222 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
223 (loadsync_<mode>): Change mode.
224 (load_quadpti, store_quadpti): New.
225 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
226 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
228 2014-04-28 Martin Jambor <mjambor@suse.cz>
230 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
231 same alias type as the original statement.
232 (subreplacement_assignment_data): New type.
233 (handle_unscalarized_data_in_subtree): New type of parameter,
234 generate new memory accesses with same alias type as the original
236 (load_assign_lhs_subreplacements): Likewise.
237 (sra_modify_constructor_assign): Generate new memory accesses with
238 same alias type as the original statement.
240 2014-04-28 Richard Biener <rguenther@suse.de>
242 * tree-pass.h (TODO_verify_il): Define.
243 (TODO_verify_all): Complete properly.
244 * passes.c (execute_function_todo): Move existing loop-closed
245 SSA verification under TODO_verify_il.
246 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
247 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
248 Fix tree sharing issue.
250 2014-04-28 Richard Biener <rguenther@suse.de>
253 * builtins.def (DEF_C11_BUILTIN): Add.
254 (BUILT_IN_ALIGNED_ALLOC): Likewise.
255 * coretypes.h (enum function_class): Add function_c11_misc.
256 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
257 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
258 (call_may_clobber_ref_p_1): Likewise.
259 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
260 (mark_all_reaching_defs_necessary_1): Likewise.
261 (propagate_necessity): Likewise.
262 (eliminate_unnecessary_stmts): Likewise.
263 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
265 2014-04-28 Richard Biener <rguenther@suse.de>
267 * tree-vrp.c (vrp_var_may_overflow): Remove.
268 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
269 with overflow immediately bump to one before that value and
270 let iteration figure out overflow status.
272 2014-04-28 Richard Biener <rguenther@suse.de>
274 * configure.ac: Do valgrind header checks unconditionally.
275 Add --enable-valgrind-annotations.
276 * system.h: Guard valgrind header inclusion with
277 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
278 * alloc-pool.c (pool_alloc, pool_free): Use
279 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
280 to guard possibly dead code.
281 * config.in: Regenerated.
282 * configure: Likewise.
284 2014-04-28 Jeff Law <law@redhat.com>
286 PR tree-optimization/60902
287 * tree-ssa-threadedge.c
288 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
289 over real defs when invalidating outputs from statements that do not
290 produce useful outputs for threading.
292 2014-04-28 Richard Biener <rguenther@suse.de>
294 PR tree-optimization/60979
295 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
296 SCOPs that end in a block with a successor with abnormal
299 2014-04-28 Richard Biener <rguenther@suse.de>
301 * tree-pass.h (execute_pass_list): Adjust prototype.
302 * passes.c (pass_manager::execute_early_local_passes):
304 (do_per_function): Change callback signature, push all actual
305 work to the callbals.
306 (do_per_function_toporder): Likewise.
307 (execute_function_dump): Adjust.
308 (execute_function_todo): Likewise.
309 (clear_last_verified): Likewise.
310 (verify_curr_properties): Likewise.
311 (update_properties_after_pass): Likewise.
312 (execute_pass_list_1): Split out from ...
313 (execute_pass_list): ... here. Adjust.
314 (execute_ipa_pass_list): Likewise.
315 * cgraphunit.c (cgraph_add_new_function): Adjust.
316 (analyze_function): Likewise.
317 (expand_function): Likewise.
318 * cgraph.c (release_function_body): Free dominance info
319 here instead of asserting it was magically freed elsewhere.
321 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
323 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
324 * configure: Regenerate.
325 * config/sparc/sparc.opt (muser-mode): New option.
326 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
328 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
329 * doc/invoke.texi (SPARC options): Document -muser-mode.
331 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
333 * cselib.c (find_slot_memmode): Delete.
334 (cselib_hasher): Change compare_type to a struct.
335 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
337 (preserve_constants_and_equivs): Adjust for new compare_type.
338 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
339 (wrap_constant): Delete.
340 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
342 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
344 * doc/install.texi (Building with profile feedback): Remove
347 2014-04-26 Tom de Vries <tom@codesourcery.com>
349 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
352 2014-04-25 Cary Coutant <ccoutant@google.com>
355 * dwarf2out.c (should_move_die_to_comdat): A type definition
356 can contain a subprogram definition, but don't move it to a
358 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
359 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
361 (clone_tree_hash): Rename to...
362 (clone_tree_partial): ...this; change callers. Copy
363 DW_TAG_subprogram DIEs as declarations.
364 (copy_decls_walk): Don't copy children of a declaration into a
367 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
370 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
373 2014-04-25 Jiong Wang <jiong.wang@arm.com>
375 * config/arm/predicates.md (call_insn_operand): Add long_call check.
376 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
378 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
381 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
383 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
386 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
388 PR tree-optimization/60930
389 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
390 creating a multiply candidate by folding two constant
391 multiplicands when the result overflows.
393 2014-04-25 Jakub Jelinek <jakub@redhat.com>
395 PR tree-optimization/60960
396 * tree-vect-generic.c (expand_vector_operation): Only call
397 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
399 2014-04-25 Tom de Vries <tom@codesourcery.com>
401 * expr.c (clobber_reg_mode): New function.
402 * expr.h (clobber_reg): New function.
404 2014-04-25 Tom de Vries <tom@codesourcery.com>
406 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
409 2014-04-25 Radovan Obradovic <robradovic@mips.com>
410 Tom de Vries <tom@codesourcery.com>
412 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
414 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
415 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
416 new argument to find_all_hard_reg_sets call.
418 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
420 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
421 Use HOST_WIDE_INT_C for mask literal.
422 (aarch_rev16_shleft_mask_imm_p): Likewise.
424 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
427 * config/sparc/sparc.md (ashlsi3_extend): Delete.
429 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
431 PR preprocessor/56540
432 * config/i386/i386-c.c (ix86_target_macros): Define
433 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
435 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
437 * configure.ac (tga_func): Remove.
438 (LIB_TLS_SPEC): Remove.
439 * configure: Regenerate.
440 * config.in: Regenerate.
441 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
443 2014-04-25 Richard Biener <rguenther@suse.de>
446 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
447 call stmt use/clobber sets during stmt walk instead of
448 walking the possibly incomplete set of caller edges.
450 2014-04-25 Richard Biener <rguenther@suse.de>
453 * passes.c (apply_ipa_transforms): Inline into only caller ...
454 (execute_one_pass): ... here. Properly bring in function
455 bodies for nodes we want to apply IPA transforms to.
457 2014-04-24 Cong Hou <congh@google.com>
459 PR tree-optimization/60896
460 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
461 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
462 (vect_mark_pattern_stmts): Set the def type of all statements in
463 PATTERN_DEF_SEQ as vect_internal_def.
465 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
467 * doc/extend.texi (PowerPC Built-in Functions): Document new
468 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
470 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
472 * config/rs6000/predicates.md (const_0_to_3_operand): New
473 predicate to match 0..3 integer constants.
475 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
476 to support adding miscellaneous builtin functions.
477 (BU_DFP_MISC_2): Likewise.
478 (BU_P7_MISC_1): Likewise.
479 (BU_P7_MISC_2): Likewise.
480 (BU_P8V_MISC_3): Likewise.
481 (BU_MISC_1): Likewise.
482 (BU_MISC_2): Likewise.
483 (DIVWE): Add extended divide builtin functions.
491 (DXEX): Add decimal floating-point builtin functions.
503 (CDTBCD): Add new BCD builtin functions.
507 (BCDADD_LT): Likewise.
508 (BCDADD_EQ): Likewise.
509 (BCDADD_GT): Likewise.
510 (BCDADD_OV): Likewise.
512 (BCDSUB_LT): Likewise.
513 (BCDSUB_EQ): Likewise.
514 (BCDSUB_GT): Likewise.
515 (BCDSUB_OV): Likewise.
516 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
517 (UNPACK_TD): Likewise.
519 (UNPACK_TF): Likewise.
520 (UNPACK_TF_0): Likewise.
521 (UNPACK_TF_1): Likewise.
522 (PACK_V1TI): Likewise.
523 (UNPACK_V1TI): Likewise.
525 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
526 support for decimal floating point builtin functions.
527 (rs6000_expand_ternop_builtin): Add checks for the new builtin
528 functions that take constant arguments.
529 (rs6000_invalid_builtin): Add decimal floating point builtin
531 (rs6000_init_builtins): Setup long double, _Decimal64, and
532 _Decimal128 types for new builtin functions.
533 (builtin_function_type): Set the unsigned flags appropriately for
534 the new builtin functions.
535 (rs6000_opt_masks): Add support for decimal floating point builtin
538 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
539 floating point builtin functions.
540 (RS6000_BTM_COMMON): Likewise.
541 (RS6000_BTI_long_double): Likewise.
542 (RS6000_BTI_dfloat64): Likewise.
543 (RS6000_BTI_dfloat128): Likewise.
544 (long_double_type_internal_node): Likewise.
545 (dfloat64_type_internal_node): Likewise.
546 (dfloat128_type_internal_node): Likewise.
548 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
549 2.07 bcd arithmetic instructions.
550 (UNSPEC_BCDSUB): Likewise.
551 (UNSPEC_BCD_OVERFLOW): Likewise.
552 (UNSPEC_BCD_ADD_SUB): Likewise.
553 (bcd_add_sub): Likewise.
554 (BCD_TEST): Likewise.
555 (bcd<bcd_add_sub>): Likewise.
556 (bcd<bcd_add_sub>_test): Likewise.
557 (bcd<bcd_add_sub>_test2): Likewise.
558 (bcd<bcd_add_sub>_<code>): Likewise.
559 (peephole2 for combined bcd ops): Likewise.
561 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
562 decimal floating point builtin functions.
563 (UNSPEC_DENBCD): Likewise.
564 (UNSPEC_DXEX): Likewise.
565 (UNSPEC_DIEX): Likewise.
566 (UNSPEC_DSCLI): Likewise.
567 (UNSPEC_DSCRI): Likewise.
568 (D64_D128): Likewise.
569 (dfp_suffix): Likewise.
570 (dfp_ddedpd_<mode>): Likewise.
571 (dfp_denbcd_<mode>): Likewise.
572 (dfp_dxex_<mode>): Likewise.
573 (dfp_diex_<mode>): Likewise.
574 (dfp_dscli_<mode>): Likewise.
575 (dfp_dscri_<mode>): Likewise.
577 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
579 (UNSPEC_CDTBCD): Likewise.
580 (UNSPEC_CBCDTD): Likewise.
581 (UNSPEC_DIVE): Add support for new extended divide builtin
583 (UNSPEC_DIVEO): Likewise.
584 (UNSPEC_DIVEU): Likewise.
585 (UNSPEC_DIVEUO): Likewise.
586 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
587 pack/unpack 128-bit types.
588 (UNSPEC_PACK_128BIT): Likewise.
589 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
590 (udiv<mode>3): Use idiv_ldiv mode attribute.
591 (div<mode>3): Likewise.
592 (addg6s): Add new BCD builtin functions.
595 (UNSPEC_DIV_EXTEND): Add support for new extended divide
597 (div_extend): Likewise.
598 (div<div_extend>_<mode>"): Likewise.
599 (FP128_64): Add support for new builtin functions to pack/unpack
601 (unpack<mode>): Likewise.
602 (unpacktf_0): Likewise.
603 (unpacktf_1): Likewise.
604 (unpack<mode>_dm): Likewise.
605 (unpack<mode>_nodm): Likewise.
606 (pack<mode>): Likewise.
607 (unpackv1ti): Likewise.
608 (packv1ti): Likewise.
610 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
612 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
615 2014-04-24 Jakub Jelinek <jakub@redhat.com>
617 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
618 * gimplify.c (omp_is_private): Change last argument's type to int.
619 Only diagnose lastprivate if the simd argument is 1, only diagnose
620 linear if the simd argument is 2.
621 (gimplify_omp_for): Adjust omp_is_private callers. When adding
622 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
623 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
624 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
625 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
626 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
627 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
628 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
629 * tree-nested.c (convert_nonlocal_omp_clauses,
630 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
632 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
635 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
638 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
640 * flag-types.h (enum ivar_visibility): Add.
642 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
644 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
647 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
649 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
651 2014-04-24 Radovan Obradovic <robradovic@mips.com>
652 Tom de Vries <tom@codesourcery.com>
654 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
655 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
657 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
658 * emit-rtl.c (try_split): Same.
660 2014-04-24 Radovan Obradovic <robradovic@mips.com>
661 Tom de Vries <tom@codesourcery.com>
663 * common.opt (fuse-caller-save): New option.
665 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
667 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
670 2014-04-24 Richard Biener <rguenther@suse.de>
672 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
673 during TER and instead use the sepops interface for expanding
674 non-GIMPLE_SINGLE_RHS.
676 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
678 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
679 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
681 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
683 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
684 assembler 64-bit option.
685 * configure: Regenerate.
687 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
689 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
690 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
691 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
692 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
693 (TARGET_CRYPTO): Take TARGET_SIMD into account.
695 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
697 * config/aarch64/aarch64-builtins.c
698 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
699 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
700 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
701 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
703 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
704 (Vrevsuff): New mode attribute.
706 2014-04-24 Terry Guo <terry.guo@arm.com>
708 * config/arm/arm.h (machine_function): Define variable
709 after_arm_reorg here.
710 * config/arm/arm.c (after_arm_reorg): Remove the definition.
711 (arm_split_constant): Update the way to access variable
714 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
716 2014-04-23 Tom de Vries <tom@codesourcery.com>
718 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
720 2014-04-23 David Malcolm <dmalcolm@redhat.com>
722 * is-a.h: Update comments to reflect the following changes to the
723 "pointerness" of the API, making the template parameter match the
724 return type, allowing use of is-a.h with typedefs of pointers.
725 (is_a_helper::cast): Return a T rather then a pointer to a T, so
726 that the return type matches the parameter to the is_a_helper.
728 (dyn_cast): Likewise.
730 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
731 pointer from the is-a.h API.
733 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
734 (is_a_helper <cgraph_node *>::test): ...this, matching change to
736 (is_a_helper <varpool_node>::test): Likewise, convert to...
737 (is_a_helper <varpool_node *>::test): ...this.
739 (varpool_first_variable): Update for removal of implicit pointer
741 (varpool_next_variable): Likewise.
742 (varpool_first_static_initializer): Likewise.
743 (varpool_next_static_initializer): Likewise.
744 (varpool_first_defined_variable): Likewise.
745 (varpool_next_defined_variable): Likewise.
746 (cgraph_first_defined_function): Likewise.
747 (cgraph_next_defined_function): Likewise.
748 (cgraph_first_function): Likewise.
749 (cgraph_next_function): Likewise.
750 (cgraph_first_function_with_gimple_body): Likewise.
751 (cgraph_next_function_with_gimple_body): Likewise.
752 (cgraph_alias_target): Likewise.
753 (varpool_alias_target): Likewise.
754 (cgraph_function_or_thunk_node): Likewise.
755 (varpool_variable_node): Likewise.
756 (symtab_real_symbol_p): Likewise.
757 * cgraphunit.c (referred_to_p): Likewise.
758 (analyze_functions): Likewise.
759 (handle_alias_pairs): Likewise.
760 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
761 * gimple-ssa.h (gimple_vuse_op): Likewise.
762 (gimple_vdef_op): Likewise.
763 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
764 * gimple.c (gimple_build_asm_1): Likewise.
765 (gimple_build_try): Likewise.
766 (gimple_build_resx): Likewise.
767 (gimple_build_eh_dispatch): Likewise.
768 (gimple_build_omp_for): Likewise.
769 (gimple_omp_for_set_clauses): Likewise.
771 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
772 (is_a_helper <gimple_statement_asm *>::test): ...this.
773 (is_a_helper <gimple_statement_bind>::test): Convert to...
774 (is_a_helper <gimple_statement_bind *>::test): ...this.
775 (is_a_helper <gimple_statement_call>::test): Convert to...
776 (is_a_helper <gimple_statement_call *>::test): ...this.
777 (is_a_helper <gimple_statement_catch>::test): Convert to...
778 (is_a_helper <gimple_statement_catch *>::test): ...this.
779 (is_a_helper <gimple_statement_resx>::test): Convert to...
780 (is_a_helper <gimple_statement_resx *>::test): ...this.
781 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
782 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
783 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
784 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
785 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
786 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
787 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
788 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
789 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
790 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
791 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
792 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
793 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
794 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
795 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
796 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
797 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
798 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
799 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
800 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
801 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
802 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
803 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
804 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
805 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
806 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
807 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
808 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
809 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
810 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
811 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
812 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
813 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
814 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
815 (is_a_helper <gimple_statement_phi>::test): Convert to...
816 (is_a_helper <gimple_statement_phi *>::test): ...this.
817 (is_a_helper <gimple_statement_transaction>::test): Convert to...
818 (is_a_helper <gimple_statement_transaction *>::test): ...this.
819 (is_a_helper <gimple_statement_try>::test): Convert to...
820 (is_a_helper <gimple_statement_try *>::test): ...this.
821 (is_a_helper <gimple_statement_wce>::test): Convert to...
822 (is_a_helper <gimple_statement_wce *>::test): ...this.
823 (is_a_helper <const gimple_statement_asm>::test): Convert to...
824 (is_a_helper <const gimple_statement_asm *>::test): ...this.
825 (is_a_helper <const gimple_statement_bind>::test): Convert to...
826 (is_a_helper <const gimple_statement_bind *>::test): ...this.
827 (is_a_helper <const gimple_statement_call>::test): Convert to...
828 (is_a_helper <const gimple_statement_call *>::test): ...this.
829 (is_a_helper <const gimple_statement_catch>::test): Convert to...
830 (is_a_helper <const gimple_statement_catch *>::test): ...this.
831 (is_a_helper <const gimple_statement_resx>::test): Convert to...
832 (is_a_helper <const gimple_statement_resx *>::test): ...this.
833 (is_a_helper <const gimple_statement_eh_dispatch>::test):
835 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
837 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
839 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
840 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
842 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
844 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
846 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
848 (is_a_helper <const gimple_statement_omp_return>::test): Convert
850 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
851 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
853 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
854 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
856 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
857 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
858 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
859 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
861 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
862 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
864 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
865 (is_a_helper <const gimple_statement_omp_target>::test): Convert
867 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
868 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
870 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
871 (is_a_helper <const gimple_statement_omp_single>::test): Convert
873 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
874 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
876 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
877 (is_a_helper <const gimple_statement_omp_task>::test): Convert
879 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
880 (is_a_helper <const gimple_statement_phi>::test): Convert to...
881 (is_a_helper <const gimple_statement_phi *>::test): ...this.
882 (is_a_helper <const gimple_statement_transaction>::test): Convert
884 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
885 (is_a_helper <const gimple_statement_with_ops>::test): Convert
887 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
888 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
889 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
890 (is_a_helper <const gimple_statement_with_memory_ops>::test):
892 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
894 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
896 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
898 (gimple_use_ops): Update for removal of implicit pointer from the
900 (gimple_set_use_ops): Likewise.
901 (gimple_vuse): Likewise.
902 (gimple_vdef): Likewise.
903 (gimple_vuse_ptr): Likewise.
904 (gimple_vdef_ptr): Likewise.
905 (gimple_set_vuse): Likewise.
906 (gimple_set_vdef): Likewise.
907 (gimple_omp_return_set_lhs): Likewise.
908 (gimple_omp_return_lhs): Likewise.
909 (gimple_omp_return_lhs_ptr): Likewise.
910 (gimple_call_fntype): Likewise.
911 (gimple_call_set_fntype): Likewise.
912 (gimple_call_set_internal_fn): Likewise.
913 (gimple_call_use_set): Likewise.
914 (gimple_call_clobber_set): Likewise.
915 (gimple_bind_vars): Likewise.
916 (gimple_bind_set_vars): Likewise.
917 (gimple_bind_body_ptr): Likewise.
918 (gimple_bind_set_body): Likewise.
919 (gimple_bind_add_stmt): Likewise.
920 (gimple_bind_block): Likewise.
921 (gimple_bind_set_block): Likewise.
922 (gimple_asm_ninputs): Likewise.
923 (gimple_asm_noutputs): Likewise.
924 (gimple_asm_nclobbers): Likewise.
925 (gimple_asm_nlabels): Likewise.
926 (gimple_asm_input_op): Likewise.
927 (gimple_asm_input_op_ptr): Likewise.
928 (gimple_asm_output_op): Likewise.
929 (gimple_asm_output_op_ptr): Likewise.
930 (gimple_asm_set_output_op): Likewise.
931 (gimple_asm_clobber_op): Likewise.
932 (gimple_asm_set_clobber_op): Likewise.
933 (gimple_asm_label_op): Likewise.
934 (gimple_asm_set_label_op): Likewise.
935 (gimple_asm_string): Likewise.
936 (gimple_catch_types): Likewise.
937 (gimple_catch_types_ptr): Likewise.
938 (gimple_catch_handler_ptr): Likewise.
939 (gimple_catch_set_types): Likewise.
940 (gimple_catch_set_handler): Likewise.
941 (gimple_eh_filter_types): Likewise.
942 (gimple_eh_filter_types_ptr): Likewise.
943 (gimple_eh_filter_failure_ptr): Likewise.
944 (gimple_eh_filter_set_types): Likewise.
945 (gimple_eh_filter_set_failure): Likewise.
946 (gimple_eh_must_not_throw_fndecl): Likewise.
947 (gimple_eh_must_not_throw_set_fndecl): Likewise.
948 (gimple_eh_else_n_body_ptr): Likewise.
949 (gimple_eh_else_e_body_ptr): Likewise.
950 (gimple_eh_else_set_n_body): Likewise.
951 (gimple_eh_else_set_e_body): Likewise.
952 (gimple_try_eval_ptr): Likewise.
953 (gimple_try_cleanup_ptr): Likewise.
954 (gimple_try_set_eval): Likewise.
955 (gimple_try_set_cleanup): Likewise.
956 (gimple_wce_cleanup_ptr): Likewise.
957 (gimple_wce_set_cleanup): Likewise.
958 (gimple_phi_capacity): Likewise.
959 (gimple_phi_num_args): Likewise.
960 (gimple_phi_result): Likewise.
961 (gimple_phi_result_ptr): Likewise.
962 (gimple_phi_set_result): Likewise.
963 (gimple_phi_arg): Likewise.
964 (gimple_phi_set_arg): Likewise.
965 (gimple_resx_region): Likewise.
966 (gimple_resx_set_region): Likewise.
967 (gimple_eh_dispatch_region): Likewise.
968 (gimple_eh_dispatch_set_region): Likewise.
969 (gimple_omp_critical_name): Likewise.
970 (gimple_omp_critical_name_ptr): Likewise.
971 (gimple_omp_critical_set_name): Likewise.
972 (gimple_omp_for_clauses): Likewise.
973 (gimple_omp_for_clauses_ptr): Likewise.
974 (gimple_omp_for_set_clauses): Likewise.
975 (gimple_omp_for_collapse): Likewise.
976 (gimple_omp_for_index): Likewise.
977 (gimple_omp_for_index_ptr): Likewise.
978 (gimple_omp_for_set_index): Likewise.
979 (gimple_omp_for_initial): Likewise.
980 (gimple_omp_for_initial_ptr): Likewise.
981 (gimple_omp_for_set_initial): Likewise.
982 (gimple_omp_for_final): Likewise.
983 (gimple_omp_for_final_ptr): Likewise.
984 (gimple_omp_for_set_final): Likewise.
985 (gimple_omp_for_incr): Likewise.
986 (gimple_omp_for_incr_ptr): Likewise.
987 (gimple_omp_for_set_incr): Likewise.
988 (gimple_omp_for_pre_body_ptr): Likewise.
989 (gimple_omp_for_set_pre_body): Likewise.
990 (gimple_omp_parallel_clauses): Likewise.
991 (gimple_omp_parallel_clauses_ptr): Likewise.
992 (gimple_omp_parallel_set_clauses): Likewise.
993 (gimple_omp_parallel_child_fn): Likewise.
994 (gimple_omp_parallel_child_fn_ptr): Likewise.
995 (gimple_omp_parallel_set_child_fn): Likewise.
996 (gimple_omp_parallel_data_arg): Likewise.
997 (gimple_omp_parallel_data_arg_ptr): Likewise.
998 (gimple_omp_parallel_set_data_arg): Likewise.
999 (gimple_omp_task_clauses): Likewise.
1000 (gimple_omp_task_clauses_ptr): Likewise.
1001 (gimple_omp_task_set_clauses): Likewise.
1002 (gimple_omp_task_child_fn): Likewise.
1003 (gimple_omp_task_child_fn_ptr): Likewise.
1004 (gimple_omp_task_set_child_fn): Likewise.
1005 (gimple_omp_task_data_arg): Likewise.
1006 (gimple_omp_task_data_arg_ptr): Likewise.
1007 (gimple_omp_task_set_data_arg): Likewise.
1008 (gimple_omp_taskreg_clauses): Likewise.
1009 (gimple_omp_taskreg_clauses_ptr): Likewise.
1010 (gimple_omp_taskreg_set_clauses): Likewise.
1011 (gimple_omp_taskreg_child_fn): Likewise.
1012 (gimple_omp_taskreg_child_fn_ptr): Likewise.
1013 (gimple_omp_taskreg_set_child_fn): Likewise.
1014 (gimple_omp_taskreg_data_arg): Likewise.
1015 (gimple_omp_taskreg_data_arg_ptr): Likewise.
1016 (gimple_omp_taskreg_set_data_arg): Likewise.
1017 (gimple_omp_task_copy_fn): Likewise.
1018 (gimple_omp_task_copy_fn_ptr): Likewise.
1019 (gimple_omp_task_set_copy_fn): Likewise.
1020 (gimple_omp_task_arg_size): Likewise.
1021 (gimple_omp_task_arg_size_ptr): Likewise.
1022 (gimple_omp_task_set_arg_size): Likewise.
1023 (gimple_omp_task_arg_align): Likewise.
1024 (gimple_omp_task_arg_align_ptr): Likewise.
1025 (gimple_omp_task_set_arg_align): Likewise.
1026 (gimple_omp_single_clauses): Likewise.
1027 (gimple_omp_single_clauses_ptr): Likewise.
1028 (gimple_omp_single_set_clauses): Likewise.
1029 (gimple_omp_target_clauses): Likewise.
1030 (gimple_omp_target_clauses_ptr): Likewise.
1031 (gimple_omp_target_set_clauses): Likewise.
1032 (gimple_omp_target_child_fn): Likewise.
1033 (gimple_omp_target_child_fn_ptr): Likewise.
1034 (gimple_omp_target_set_child_fn): Likewise.
1035 (gimple_omp_target_data_arg): Likewise.
1036 (gimple_omp_target_data_arg_ptr): Likewise.
1037 (gimple_omp_target_set_data_arg): Likewise.
1038 (gimple_omp_teams_clauses): Likewise.
1039 (gimple_omp_teams_clauses_ptr): Likewise.
1040 (gimple_omp_teams_set_clauses): Likewise.
1041 (gimple_omp_sections_clauses): Likewise.
1042 (gimple_omp_sections_clauses_ptr): Likewise.
1043 (gimple_omp_sections_set_clauses): Likewise.
1044 (gimple_omp_sections_control): Likewise.
1045 (gimple_omp_sections_control_ptr): Likewise.
1046 (gimple_omp_sections_set_control): Likewise.
1047 (gimple_omp_for_set_cond): Likewise.
1048 (gimple_omp_for_cond): Likewise.
1049 (gimple_omp_atomic_store_set_val): Likewise.
1050 (gimple_omp_atomic_store_val): Likewise.
1051 (gimple_omp_atomic_store_val_ptr): Likewise.
1052 (gimple_omp_atomic_load_set_lhs): Likewise.
1053 (gimple_omp_atomic_load_lhs): Likewise.
1054 (gimple_omp_atomic_load_lhs_ptr): Likewise.
1055 (gimple_omp_atomic_load_set_rhs): Likewise.
1056 (gimple_omp_atomic_load_rhs): Likewise.
1057 (gimple_omp_atomic_load_rhs_ptr): Likewise.
1058 (gimple_omp_continue_control_def): Likewise.
1059 (gimple_omp_continue_control_def_ptr): Likewise.
1060 (gimple_omp_continue_set_control_def): Likewise.
1061 (gimple_omp_continue_control_use): Likewise.
1062 (gimple_omp_continue_control_use_ptr): Likewise.
1063 (gimple_omp_continue_set_control_use): Likewise.
1064 (gimple_transaction_body_ptr): Likewise.
1065 (gimple_transaction_label): Likewise.
1066 (gimple_transaction_label_ptr): Likewise.
1067 (gimple_transaction_set_body): Likewise.
1068 (gimple_transaction_set_label): Likewise.
1070 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
1071 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1072 * ipa-ref.c (ipa_record_reference): Likewise.
1073 * ipa-reference.c (analyze_function): Likewise.
1074 (ipa_reference_write_optimization_summary): Likewise.
1075 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
1076 (address_taken_from_non_vtable_p): Likewise.
1077 (comdat_can_be_unshared_p_1): Likewise.
1078 * lto-cgraph.c (lto_output_ref): Likewise.
1079 (add_references): Likewise.
1080 (compute_ltrans_boundary): Likewise.
1081 (output_symtab): Likewise.
1082 (input_ref): Likewise.
1083 (input_cgraph_1): Likewise.
1084 (output_cgraph_opt_summary): Likewise.
1085 * lto-streamer-out.c (lto_output): Likewise.
1086 (output_symbol_p): Likewise.
1087 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
1088 (lsei_start_function_in_partition): Likewise.
1089 (lsei_next_variable_in_partition): Likewise.
1090 (lsei_start_variable_in_partition): Likewise.
1091 * symtab.c (insert_to_assembler_name_hash): Likewise.
1092 (unlink_from_assembler_name_hash): Likewise.
1093 (symtab_unregister_node): Likewise.
1094 (symtab_remove_node): Likewise.
1095 (dump_symtab_node): Likewise.
1096 (verify_symtab_base): Likewise.
1097 (verify_symtab_node): Likewise.
1098 (symtab_make_decl_local): Likewise.
1099 (symtab_alias_ultimate_target): Likewise.
1100 (symtab_resolve_alias): Likewise.
1101 (symtab_get_symbol_partitioning_class): Likewise.
1102 * tree-phinodes.c (allocate_phi_node): Likewise.
1103 (reserve_phi_args_for_new_edge): Likewise.
1104 (remove_phi_args): Likewise.
1105 * varpool.c (varpool_node_for_asm): Likewise.
1106 (varpool_remove_unreferenced_decls): Likewise.
1108 2014-04-23 Jeff Law <law@redhat.com>
1110 PR tree-optimization/60902
1111 * tree-ssa-threadedge.c
1112 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
1113 invalidate outputs from statements that do not produce useful
1114 outputs for threading.
1116 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
1118 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
1119 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
1120 machine descriptions for Stack Smashing Protector.
1122 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
1124 * aarch64.md (<optab>_rol<mode>3): New pattern.
1125 (<optab>_rolsi3_uxtw): Likewise.
1126 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
1128 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
1130 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
1131 (arm_cortex_a12_tune): Likewise.
1133 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1135 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
1137 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1139 * config/arm/arm.md (arm_rev16si2): New pattern.
1140 (arm_rev16si2_alt): Likewise.
1141 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
1143 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1145 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
1146 (rev16<mode>2_alt): Likewise.
1147 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
1148 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
1149 (aarch_rev16_shleft_mask_imm_p): Likewise.
1150 (aarch_rev16_p_1): Likewise.
1151 (aarch_rev16_p): Likewise.
1152 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
1153 (aarch_rev16_shright_mask_imm_p): Likewise.
1154 (aarch_rev16_shleft_mask_imm_p): Likewise.
1156 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1158 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
1159 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
1161 (cortex_a53_extra_costs): Likewise.
1162 (cortex_a57_extra_costs): Likewise.
1163 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
1164 (cortexa7_extra_costs): Likewise.
1165 (cortexa8_extra_costs): Likewise.
1166 (cortexa12_extra_costs): Likewise.
1167 (cortexa15_extra_costs): Likewise.
1168 (v7m_extra_costs): Likewise.
1169 (arm_new_rtx_costs): Handle BSWAP.
1171 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1173 * config/arm/arm.c (cortexa8_extra_costs): New table.
1174 (arm_cortex_a8_tune): New tuning struct.
1175 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
1177 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1179 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
1181 2014-04-23 Richard Biener <rguenther@suse.de>
1183 * Makefile.in (OBJS): Remove loop-unswitch.o.
1184 * loop-unswitch.c: Delete.
1185 * tree-pass.h (make_pass_rtl_unswitch): Remove.
1186 * passes.def (pass_rtl_unswitch): Likewise.
1187 * loop-init.c (gate_rtl_unswitch): Likewise.
1188 (rtl_unswitch): Likewise.
1189 (pass_data_rtl_unswitch): Likewise.
1190 (pass_rtl_unswitch): Likewise.
1191 (make_pass_rtl_unswitch): Likewise.
1192 * rtl.h (reversed_condition): Likewise.
1193 (compare_and_jump_seq): Likewise.
1194 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
1196 * loop-unroll.c (compare_and_jump_seq): Likewise.
1198 2014-04-23 Richard Biener <rguenther@suse.de>
1200 PR tree-optimization/60903
1201 * tree-ssa-loop-im.c (analyze_memory_references): Remove
1202 commented code block.
1203 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
1204 loop flags to newly created BBs and edges.
1206 2014-04-23 Nick Clifton <nickc@redhat.com>
1208 * config/msp430/msp430.c (msp430_handle_option): Move function
1210 (msp430_option_override): Simplify mcu and mcpu option handling.
1211 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
1212 support for -mhwmult command line option.
1213 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
1214 -mhwmult command line option.
1215 (msp430_hwmult_enabled): Delete.
1216 (msp43o_output_labelref): Add support for -mhwmult command line
1218 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
1219 (umulsidi3): Likewise.
1220 * config/msp430/msp430.opt (mmcu): Add Report attribute.
1221 (mcpu, mlarge, msmall): Likewise.
1222 (mhwmult): New option.
1223 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
1225 (msp430_is_f5_mcu): Remove prototype.
1226 (msp430_use_f5_series_hwmult): Add prototype.
1227 * config/msp430/msp430-opts.h: New file.
1228 * common/config/msp430: New directory.
1229 * common/config/msp430/msp430-common.c: New file.
1230 * config.gcc (msp430): Remove target_has_targetm_common.
1231 * doc/invoke.texi: Document -mhwmult command line option.
1233 2014-04-23 Nick Clifton <nickc@redhat.com>
1235 * config/i386/cygwin.h (ENDFILE_SPEC): Include
1236 default-manifest.o if it can be found in the search path.
1237 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
1239 2014-04-23 Terry Guo <terry.guo@arm.com>
1241 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
1243 2014-04-23 Richard Biener <rguenther@suse.de>
1246 * tree-inline.c (declare_return_variable): Use mark_addressable.
1248 2014-04-23 Richard Biener <rguenther@suse.de>
1251 * loop-init.c (loop_optimizer_init): Make sure to apply
1252 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
1254 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1257 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
1259 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
1260 if flag_sanitize_undefined_trap_on_error.
1261 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
1262 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
1263 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1264 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1265 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1266 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1267 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1268 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1269 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1270 * ubsan.c (ubsan_instrument_unreachable): Return
1271 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1272 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1273 if flag_sanitize_undefined_trap_on_error and
1274 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1275 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1276 instrument_bool_enum_load): Emit __builtin_trap () if
1277 flag_sanitize_undefined_trap_on_error and
1278 __builtin_handle_*_abort () if !flag_sanitize_recover.
1279 * doc/invoke.texi (-fsanitize-recover,
1280 -fsanitize-undefined-trap-on-error): Document.
1282 2014-04-22 Christian Bruel <christian.bruel@st.com>
1284 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1285 Force immediates to SImode.
1287 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1289 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1291 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1292 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1293 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1294 (nios2_fpu_insn): Add entry for round.
1295 (N2FPU_NO_ERRNO_P): Define.
1296 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1298 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1300 2014-04-22 Richard Henderson <rth@redhat.com>
1302 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1303 (add<GPI>3_compare0): Remove leading * from name.
1304 (add<GPI>3_carryin): Likewise.
1305 (sub<GPI>3_compare0): Likewise.
1306 (sub<GPI>3_carryin): Likewise.
1307 (<su_optab>mulditi3): New expander.
1308 (multi3): New expander.
1309 (madd<GPI>): Remove leading * from name.
1311 2014-04-22 Martin Jambor <mjambor@suse.cz>
1313 * cgraphclones.c (cgraph_function_versioning): Copy
1314 ipa_transforms_to_apply instead of asserting it is empty.
1316 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1319 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1320 on count_exp to get mode.
1322 2014-04-22 Andrew Pinski <apinski@cavium.com>
1324 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1325 Handle TLS for ILP32.
1326 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1327 (tlsie_small_<mode>): this and handle PTR.
1328 (tlsie_small_sidi): New pattern.
1329 (tlsle_small): Change to an expand to handle ILP32.
1330 (tlsle_small_<mode>): New pattern.
1331 (tlsdesc_small): Rename to ...
1332 (tlsdesc_small_<mode>): this and handle PTR.
1334 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1336 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1338 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1340 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1341 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1342 (aarch64_types_signed_poly_qualifiers): Likewise.
1343 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1344 (aarch64_types_poly_signed_qualifiers): Likewise.
1345 (TYPES_REINTERP_SS): Type macro added.
1346 (TYPES_REINTERP_SU): Likewise.
1347 (TYPES_REINTERP_SP): Likewise.
1348 (TYPES_REINTERP_US): Likewise.
1349 (TYPES_REINTERP_PS): Likewise.
1350 (aarch64_fold_builtin): New expression folding added.
1351 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1352 Declarations removed.
1353 (REINTERP_SS): Declarations added.
1354 (REINTERP_US): Likewise.
1355 (REINTERP_PS): Likewise.
1356 (REINTERP_SU): Likewise.
1357 (REINTERP_SP): Likewise.
1358 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1359 (vreinterpretq_p8_f64): Likewise.
1360 (vreinterpret_p16_f64): Likewise.
1361 (vreinterpretq_p16_f64): Likewise.
1362 (vreinterpret_f32_f64): Likewise.
1363 (vreinterpretq_f32_f64): Likewise.
1364 (vreinterpret_f64_f32): Likewise.
1365 (vreinterpret_f64_p8): Likewise.
1366 (vreinterpret_f64_p16): Likewise.
1367 (vreinterpret_f64_s8): Likewise.
1368 (vreinterpret_f64_s16): Likewise.
1369 (vreinterpret_f64_s32): Likewise.
1370 (vreinterpret_f64_s64): Likewise.
1371 (vreinterpret_f64_u8): Likewise.
1372 (vreinterpret_f64_u16): Likewise.
1373 (vreinterpret_f64_u32): Likewise.
1374 (vreinterpret_f64_u64): Likewise.
1375 (vreinterpretq_f64_f32): Likewise.
1376 (vreinterpretq_f64_p8): Likewise.
1377 (vreinterpretq_f64_p16): Likewise.
1378 (vreinterpretq_f64_s8): Likewise.
1379 (vreinterpretq_f64_s16): Likewise.
1380 (vreinterpretq_f64_s32): Likewise.
1381 (vreinterpretq_f64_s64): Likewise.
1382 (vreinterpretq_f64_u8): Likewise.
1383 (vreinterpretq_f64_u16): Likewise.
1384 (vreinterpretq_f64_u32): Likewise.
1385 (vreinterpretq_f64_u64): Likewise.
1386 (vreinterpret_s64_f64): Likewise.
1387 (vreinterpretq_s64_f64): Likewise.
1388 (vreinterpret_u64_f64): Likewise.
1389 (vreinterpretq_u64_f64): Likewise.
1390 (vreinterpret_s8_f64): Likewise.
1391 (vreinterpretq_s8_f64): Likewise.
1392 (vreinterpret_s16_f64): Likewise.
1393 (vreinterpretq_s16_f64): Likewise.
1394 (vreinterpret_s32_f64): Likewise.
1395 (vreinterpretq_s32_f64): Likewise.
1396 (vreinterpret_u8_f64): Likewise.
1397 (vreinterpretq_u8_f64): Likewise.
1398 (vreinterpret_u16_f64): Likewise.
1399 (vreinterpretq_u16_f64): Likewise.
1400 (vreinterpret_u32_f64): Likewise.
1401 (vreinterpretq_u32_f64): Likewise.
1403 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1405 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1406 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1407 (vreinterpret_p8_s8): Likewise.
1408 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1409 (vreinterpret_p8_s16): Likewise.
1410 (vreinterpret_p8_s32): Likewise.
1411 (vreinterpret_p8_s64): Likewise.
1412 (vreinterpret_p8_f32): Likewise.
1413 (vreinterpret_p8_u8): Likewise.
1414 (vreinterpret_p8_u16): Likewise.
1415 (vreinterpret_p8_u32): Likewise.
1416 (vreinterpret_p8_u64): Likewise.
1417 (vreinterpret_p8_p16): Likewise.
1418 (vreinterpretq_p8_s8): Likewise.
1419 (vreinterpretq_p8_s16): Likewise.
1420 (vreinterpretq_p8_s32): Likewise.
1421 (vreinterpretq_p8_s64): Likewise.
1422 (vreinterpretq_p8_f32): Likewise.
1423 (vreinterpretq_p8_u8): Likewise.
1424 (vreinterpretq_p8_u16): Likewise.
1425 (vreinterpretq_p8_u32): Likewise.
1426 (vreinterpretq_p8_u64): Likewise.
1427 (vreinterpretq_p8_p16): Likewise.
1428 (vreinterpret_p16_s8): Likewise.
1429 (vreinterpret_p16_s16): Likewise.
1430 (vreinterpret_p16_s32): Likewise.
1431 (vreinterpret_p16_s64): Likewise.
1432 (vreinterpret_p16_f32): Likewise.
1433 (vreinterpret_p16_u8): Likewise.
1434 (vreinterpret_p16_u16): Likewise.
1435 (vreinterpret_p16_u32): Likewise.
1436 (vreinterpret_p16_u64): Likewise.
1437 (vreinterpret_p16_p8): Likewise.
1438 (vreinterpretq_p16_s8): Likewise.
1439 (vreinterpretq_p16_s16): Likewise.
1440 (vreinterpretq_p16_s32): Likewise.
1441 (vreinterpretq_p16_s64): Likewise.
1442 (vreinterpretq_p16_f32): Likewise.
1443 (vreinterpretq_p16_u8): Likewise.
1444 (vreinterpretq_p16_u16): Likewise.
1445 (vreinterpretq_p16_u32): Likewise.
1446 (vreinterpretq_p16_u64): Likewise.
1447 (vreinterpretq_p16_p8): Likewise.
1448 (vreinterpret_f32_s8): Likewise.
1449 (vreinterpret_f32_s16): Likewise.
1450 (vreinterpret_f32_s32): Likewise.
1451 (vreinterpret_f32_s64): Likewise.
1452 (vreinterpret_f32_u8): Likewise.
1453 (vreinterpret_f32_u16): Likewise.
1454 (vreinterpret_f32_u32): Likewise.
1455 (vreinterpret_f32_u64): Likewise.
1456 (vreinterpret_f32_p8): Likewise.
1457 (vreinterpret_f32_p16): Likewise.
1458 (vreinterpretq_f32_s8): Likewise.
1459 (vreinterpretq_f32_s16): Likewise.
1460 (vreinterpretq_f32_s32): Likewise.
1461 (vreinterpretq_f32_s64): Likewise.
1462 (vreinterpretq_f32_u8): Likewise.
1463 (vreinterpretq_f32_u16): Likewise.
1464 (vreinterpretq_f32_u32): Likewise.
1465 (vreinterpretq_f32_u64): Likewise.
1466 (vreinterpretq_f32_p8): Likewise.
1467 (vreinterpretq_f32_p16): Likewise.
1468 (vreinterpret_s64_s8): Likewise.
1469 (vreinterpret_s64_s16): Likewise.
1470 (vreinterpret_s64_s32): Likewise.
1471 (vreinterpret_s64_f32): Likewise.
1472 (vreinterpret_s64_u8): Likewise.
1473 (vreinterpret_s64_u16): Likewise.
1474 (vreinterpret_s64_u32): Likewise.
1475 (vreinterpret_s64_u64): Likewise.
1476 (vreinterpret_s64_p8): Likewise.
1477 (vreinterpret_s64_p16): Likewise.
1478 (vreinterpretq_s64_s8): Likewise.
1479 (vreinterpretq_s64_s16): Likewise.
1480 (vreinterpretq_s64_s32): Likewise.
1481 (vreinterpretq_s64_f32): Likewise.
1482 (vreinterpretq_s64_u8): Likewise.
1483 (vreinterpretq_s64_u16): Likewise.
1484 (vreinterpretq_s64_u32): Likewise.
1485 (vreinterpretq_s64_u64): Likewise.
1486 (vreinterpretq_s64_p8): Likewise.
1487 (vreinterpretq_s64_p16): Likewise.
1488 (vreinterpret_u64_s8): Likewise.
1489 (vreinterpret_u64_s16): Likewise.
1490 (vreinterpret_u64_s32): Likewise.
1491 (vreinterpret_u64_s64): Likewise.
1492 (vreinterpret_u64_f32): Likewise.
1493 (vreinterpret_u64_u8): Likewise.
1494 (vreinterpret_u64_u16): Likewise.
1495 (vreinterpret_u64_u32): Likewise.
1496 (vreinterpret_u64_p8): Likewise.
1497 (vreinterpret_u64_p16): Likewise.
1498 (vreinterpretq_u64_s8): Likewise.
1499 (vreinterpretq_u64_s16): Likewise.
1500 (vreinterpretq_u64_s32): Likewise.
1501 (vreinterpretq_u64_s64): Likewise.
1502 (vreinterpretq_u64_f32): Likewise.
1503 (vreinterpretq_u64_u8): Likewise.
1504 (vreinterpretq_u64_u16): Likewise.
1505 (vreinterpretq_u64_u32): Likewise.
1506 (vreinterpretq_u64_p8): Likewise.
1507 (vreinterpretq_u64_p16): Likewise.
1508 (vreinterpret_s8_s16): Likewise.
1509 (vreinterpret_s8_s32): Likewise.
1510 (vreinterpret_s8_s64): Likewise.
1511 (vreinterpret_s8_f32): Likewise.
1512 (vreinterpret_s8_u8): Likewise.
1513 (vreinterpret_s8_u16): Likewise.
1514 (vreinterpret_s8_u32): Likewise.
1515 (vreinterpret_s8_u64): Likewise.
1516 (vreinterpret_s8_p8): Likewise.
1517 (vreinterpret_s8_p16): Likewise.
1518 (vreinterpretq_s8_s16): Likewise.
1519 (vreinterpretq_s8_s32): Likewise.
1520 (vreinterpretq_s8_s64): Likewise.
1521 (vreinterpretq_s8_f32): Likewise.
1522 (vreinterpretq_s8_u8): Likewise.
1523 (vreinterpretq_s8_u16): Likewise.
1524 (vreinterpretq_s8_u32): Likewise.
1525 (vreinterpretq_s8_u64): Likewise.
1526 (vreinterpretq_s8_p8): Likewise.
1527 (vreinterpretq_s8_p16): Likewise.
1528 (vreinterpret_s16_s8): Likewise.
1529 (vreinterpret_s16_s32): Likewise.
1530 (vreinterpret_s16_s64): Likewise.
1531 (vreinterpret_s16_f32): Likewise.
1532 (vreinterpret_s16_u8): Likewise.
1533 (vreinterpret_s16_u16): Likewise.
1534 (vreinterpret_s16_u32): Likewise.
1535 (vreinterpret_s16_u64): Likewise.
1536 (vreinterpret_s16_p8): Likewise.
1537 (vreinterpret_s16_p16): Likewise.
1538 (vreinterpretq_s16_s8): Likewise.
1539 (vreinterpretq_s16_s32): Likewise.
1540 (vreinterpretq_s16_s64): Likewise.
1541 (vreinterpretq_s16_f32): Likewise.
1542 (vreinterpretq_s16_u8): Likewise.
1543 (vreinterpretq_s16_u16): Likewise.
1544 (vreinterpretq_s16_u32): Likewise.
1545 (vreinterpretq_s16_u64): Likewise.
1546 (vreinterpretq_s16_p8): Likewise.
1547 (vreinterpretq_s16_p16): Likewise.
1548 (vreinterpret_s32_s8): Likewise.
1549 (vreinterpret_s32_s16): Likewise.
1550 (vreinterpret_s32_s64): Likewise.
1551 (vreinterpret_s32_f32): Likewise.
1552 (vreinterpret_s32_u8): Likewise.
1553 (vreinterpret_s32_u16): Likewise.
1554 (vreinterpret_s32_u32): Likewise.
1555 (vreinterpret_s32_u64): Likewise.
1556 (vreinterpret_s32_p8): Likewise.
1557 (vreinterpret_s32_p16): Likewise.
1558 (vreinterpretq_s32_s8): Likewise.
1559 (vreinterpretq_s32_s16): Likewise.
1560 (vreinterpretq_s32_s64): Likewise.
1561 (vreinterpretq_s32_f32): Likewise.
1562 (vreinterpretq_s32_u8): Likewise.
1563 (vreinterpretq_s32_u16): Likewise.
1564 (vreinterpretq_s32_u32): Likewise.
1565 (vreinterpretq_s32_u64): Likewise.
1566 (vreinterpretq_s32_p8): Likewise.
1567 (vreinterpretq_s32_p16): Likewise.
1568 (vreinterpret_u8_s8): Likewise.
1569 (vreinterpret_u8_s16): Likewise.
1570 (vreinterpret_u8_s32): Likewise.
1571 (vreinterpret_u8_s64): Likewise.
1572 (vreinterpret_u8_f32): Likewise.
1573 (vreinterpret_u8_u16): Likewise.
1574 (vreinterpret_u8_u32): Likewise.
1575 (vreinterpret_u8_u64): Likewise.
1576 (vreinterpret_u8_p8): Likewise.
1577 (vreinterpret_u8_p16): Likewise.
1578 (vreinterpretq_u8_s8): Likewise.
1579 (vreinterpretq_u8_s16): Likewise.
1580 (vreinterpretq_u8_s32): Likewise.
1581 (vreinterpretq_u8_s64): Likewise.
1582 (vreinterpretq_u8_f32): Likewise.
1583 (vreinterpretq_u8_u16): Likewise.
1584 (vreinterpretq_u8_u32): Likewise.
1585 (vreinterpretq_u8_u64): Likewise.
1586 (vreinterpretq_u8_p8): Likewise.
1587 (vreinterpretq_u8_p16): Likewise.
1588 (vreinterpret_u16_s8): Likewise.
1589 (vreinterpret_u16_s16): Likewise.
1590 (vreinterpret_u16_s32): Likewise.
1591 (vreinterpret_u16_s64): Likewise.
1592 (vreinterpret_u16_f32): Likewise.
1593 (vreinterpret_u16_u8): Likewise.
1594 (vreinterpret_u16_u32): Likewise.
1595 (vreinterpret_u16_u64): Likewise.
1596 (vreinterpret_u16_p8): Likewise.
1597 (vreinterpret_u16_p16): Likewise.
1598 (vreinterpretq_u16_s8): Likewise.
1599 (vreinterpretq_u16_s16): Likewise.
1600 (vreinterpretq_u16_s32): Likewise.
1601 (vreinterpretq_u16_s64): Likewise.
1602 (vreinterpretq_u16_f32): Likewise.
1603 (vreinterpretq_u16_u8): Likewise.
1604 (vreinterpretq_u16_u32): Likewise.
1605 (vreinterpretq_u16_u64): Likewise.
1606 (vreinterpretq_u16_p8): Likewise.
1607 (vreinterpretq_u16_p16): Likewise.
1608 (vreinterpret_u32_s8): Likewise.
1609 (vreinterpret_u32_s16): Likewise.
1610 (vreinterpret_u32_s32): Likewise.
1611 (vreinterpret_u32_s64): Likewise.
1612 (vreinterpret_u32_f32): Likewise.
1613 (vreinterpret_u32_u8): Likewise.
1614 (vreinterpret_u32_u16): Likewise.
1615 (vreinterpret_u32_u64): Likewise.
1616 (vreinterpret_u32_p8): Likewise.
1617 (vreinterpret_u32_p16): Likewise.
1618 (vreinterpretq_u32_s8): Likewise.
1619 (vreinterpretq_u32_s16): Likewise.
1620 (vreinterpretq_u32_s32): Likewise.
1621 (vreinterpretq_u32_s64): Likewise.
1622 (vreinterpretq_u32_f32): Likewise.
1623 (vreinterpretq_u32_u8): Likewise.
1624 (vreinterpretq_u32_u16): Likewise.
1625 (vreinterpretq_u32_u64): Likewise.
1626 (vreinterpretq_u32_p8): Likewise.
1627 (vreinterpretq_u32_p16): Likewise.
1629 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1631 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1633 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1636 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1637 (vqnegd_s64): Likewise.
1638 (vqabs_s64): Likewise.
1639 (vqabsd_s64): Likewise.
1641 2014-04-22 Richard Henderson <rth@redhat.com>
1643 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1644 computation to the top of the loop.
1646 2014-04-22 Renlin <renlin.li@arm.com>
1647 Jiong Wang <jiong.wang@arm.com>
1649 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1650 * config/aarch64/aarch64.c (aarch64_layout_frame)
1651 (aarch64_initial_elimination_offset): Likewise.
1653 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1655 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1658 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1660 * machmode.h (bitwise_mode_for_mode): Declare.
1661 * stor-layout.h (bitwise_type_for_mode): Likewise.
1662 * stor-layout.c (bitwise_mode_for_mode): New function.
1663 (bitwise_type_for_mode): Likewise.
1664 * builtins.c (fold_builtin_memory_op): Use it instead of
1665 int_mode_for_mode and build_nonstandard_integer_type.
1667 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1669 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1670 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1671 (*-*-solaris2*): Simplify.
1672 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1673 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1674 *-*-solaris2.9* handling.
1676 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1678 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1679 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1681 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1682 * configure: Regenerate.
1684 * config/i386/sol2-9.h: Remove.
1686 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1687 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1688 Remove Solaris 9 references.
1690 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1692 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1693 (floatuns<GPI:mode><GPF:mode>2): Remove.
1694 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1695 and floatuns conversions.
1696 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1697 and floatuns conversions.
1698 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1699 (w1,w2): New mode attributes for inequal width conversions.
1701 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1703 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1704 the output asm format.
1706 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1708 * config/aarch64/aarch64-simd.md
1709 (aarch64_cm<optab>di): Always split.
1710 (*aarch64_cm<optab>di): New.
1711 (aarch64_cmtstdi): Always split.
1712 (*aarch64_cmtstdi): New.
1714 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1716 PR tree-optimization/60823
1717 * omp-low.c (ipa_simd_modify_function_body): Go through
1718 all SSA_NAMEs and for those refering to vector arguments
1719 which are going to be replaced adjust SSA_NAME_VAR and,
1720 if it is a default definition, change it into a non-default
1721 definition assigned at the beginning of function from new_decl.
1722 (ipa_simd_modify_stmt_ops): Rewritten.
1723 * tree-dfa.c (set_ssa_default_def): When removing default def,
1724 check for NULL loc instead of NULL *loc.
1726 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1728 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1729 restrictions on core registers for DImode values in Thumb2.
1731 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1733 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1734 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1736 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1738 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1739 (*iordi_notzesidi_di): Likewise.
1740 (*iordi_notsesidi_di): Likewise.
1742 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1744 * config/arm/arm-protos.h (tune_params): New struct members.
1745 * config/arm/arm.c: Initialise tune_params per processor.
1746 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1747 for speed, based on new tune_params.
1749 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1751 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1753 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1755 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1757 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1758 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1759 (vrnda_f64): Likewise.
1760 (vrndi_f64): Likewise.
1761 (vrndm_f64): Likewise.
1762 (vrndn_f64): Likewise.
1763 (vrndp_f64): Likewise.
1764 (vrndx_f64): Likewise.
1766 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1768 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1769 GET_MODE_SIZE argument is enum machine_mode.
1771 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1774 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1775 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1777 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1780 * asan.c (asan_emit_stack_protection): Force the base to align to
1781 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1782 appropriate bits if STRICT_ALIGNMENT.
1783 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1785 (expand_used_vars): Leave a space in the stack frame for alignment
1786 if STRICT_ALIGNMENT.
1788 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1790 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1792 (gimple_store_p): Likewise.
1793 (gimple_assign_load_p): Likewise.
1794 (gimple_assign_cast_p): Likewise.
1795 (gimple_clobber_p): Likewise.
1797 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1798 rather than a gimple.
1799 (gimple_assign_cast_p): Likewise.
1801 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1804 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1805 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1807 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1808 more debug information for E500 if -mdebug=reg.
1810 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1813 * config/i386/i386.c (ix86_expand_builtin)
1814 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1815 register for target RTX.
1816 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1818 2014-04-18 Cong Hou <congh@google.com>
1820 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1821 the widen-mult pattern by handling two operands with different sizes,
1822 and operands whose size is smaller than half of the result type.
1824 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1826 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1827 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1828 (do_estimate_edge_time): Compute it.
1829 * ipa-inline.c (want_inline_small_function_p): Bypass
1830 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1832 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1834 * ipa-inline.c (spec_rem): New static variable.
1835 (dump_overall_stats): New function.
1836 (dump_inline_stats): New function.
1838 2014-04-18 Richard Henderson <rth@redhat.com>
1840 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1841 to GET_MODE_SIZE, not a reg_class_t.
1843 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1845 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1846 (vsx_xxmrglw_<mode>): Likewise.
1848 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1851 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1852 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1853 (rs6000_init_hard_regno_mode_ok): Likewise.
1855 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1857 * ipa-inline.c (inline_small_functions): Account only non-cold
1859 * doc/invoke.texi (inline-unit-growth): Update documentation.
1861 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1863 * config/rs6000/rs6000.md (addti3, subti3): New.
1865 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1868 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1869 comment. Check optimize_insn_for_size_p instead of
1870 optimize_insn_for_speed_p.
1872 2014-04-17 Martin Jambor <mjambor@suse.cz>
1874 * gimple-iterator.c (gsi_start_edge): New function.
1875 * gimple-iterator.h (gsi_start_edge): Declare.
1876 * tree-sra.c (single_non_eh_succ): New function.
1877 (disqualify_ops_if_throwing_stmt): Renamed to
1878 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1879 having one non-EH successor BB.
1880 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1881 generate loads into replacements.
1882 (sra_modify_assign): Likewise and and also use the simple path for
1884 (sra_modify_function_body): Commit statements on edges.
1886 2014-04-17 Richard Biener <rguenther@suse.de>
1889 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1890 comparison results and add clarifying comment.
1892 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1894 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1895 (blank_mode): Initialize it.
1896 (emit_mode_size_inline, emit_mode_nunits_inline,
1897 emit_mode_inner_inline): New functions.
1898 (emit_insn_modes_h): Call them and surround their output with
1899 #if GCC_VERSION >= 4001 ... #endif.
1900 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1901 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1902 mode_* arrays if the argument is __builtin_constant_p.
1903 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1904 is enum machine_mode.
1906 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1908 * passes.c (opt_pass::execute): Adjust.
1909 (pass_manager::execute_pass_mode_switching): Likewise.
1910 (early_local_passes::execute): Likewise.
1911 (execute_one_pass): Pass cfun to the pass's execute method.
1912 * tree-pass.h (opt_pass::execute): Add function * argument.
1913 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1914 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1915 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1916 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1917 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1918 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1919 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1920 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1921 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1922 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1923 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1924 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1925 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1926 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1927 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1928 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1929 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1930 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1931 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1932 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1933 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1934 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1935 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1936 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1937 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1938 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1939 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1940 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1941 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1944 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1946 * passes.c (opt_pass::gate): Take function * argument.
1947 (gate_all_early_local_passes): Merge into
1948 (early_local_passes::gate): this.
1949 (gate_all_early_optimizations): Merge into
1950 (all_early_optimizations::gate): this.
1951 (gate_all_optimizations): Mege into
1952 (all_optimizations::gate): this.
1953 (gate_all_optimizations_g): Merge into
1954 (all_optimizations_g::gate): this.
1955 (gate_rest_of_compilation): Mege into
1956 (rest_of_compilation::gate): this.
1957 (gate_postreload): Merge into
1958 (postreload::gate): this.
1959 (dump_one_pass): Pass cfun to the pass's gate method.
1960 (execute_ipa_summary_passes): Likewise.
1961 (execute_one_pass): Likewise.
1962 (ipa_write_summaries_2): Likewise.
1963 (ipa_write_optimization_summaries_1): Likewise.
1964 (ipa_read_summaries_1): Likewise.
1965 (ipa_read_optimization_summaries_1): Likewise.
1966 (execute_ipa_stmt_fixups): Likewise.
1967 * tree-pass.h (opt_pass::gate): Add function * argument.
1968 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
1969 combine-stack-adj.c, combine.c, compare-elim.c,
1970 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1971 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
1972 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
1973 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
1974 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1975 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1976 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1977 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
1978 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
1979 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
1980 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1981 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1982 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
1983 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1984 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1985 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1986 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1987 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1988 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1989 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1990 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1991 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1992 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1993 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
1994 var-tracking.c, vtable-verify.c, web.c: Adjust.
1996 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1998 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
1999 * configure: Regenerate.
2001 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2003 * passes.c (dump_one_pass): don't check pass->has_gate.
2004 (execute_ipa_summary_passes): Likewise.
2005 (execute_one_pass): Likewise.
2006 (ipa_write_summaries_2): Likewise.
2007 (ipa_write_optimization_summaries_1): Likewise.
2008 (ipa_read_optimization_summaries_1): Likewise.
2009 (execute_ipa_stmt_fixups): Likewise.
2010 * tree-pass.h (pass_data::has_gate): Remove.
2011 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
2012 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
2013 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
2014 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2015 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
2016 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
2017 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
2018 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
2019 gimple-low.c, gimple-ssa-isolate-paths.c,
2020 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
2021 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
2022 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
2023 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
2024 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
2025 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
2026 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
2027 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
2028 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
2029 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
2030 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2031 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2032 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2033 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2034 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2035 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2036 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2037 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2038 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2039 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
2040 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
2041 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
2044 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2046 * pass_manager.h (pass_manager::register_dump_files_1): Remove
2048 * passes.c (pass_manager::register_dump_files_1): Merge into
2049 (pass_manager::register_dump_files): this, and remove its handling of
2050 properties since the pass always has the properties anyway.
2051 (pass_manager::pass_manager): Adjust.
2053 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2055 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
2056 * passes.c (pass_manager::register_dump_files_1): Remove dead code
2057 dealing with properties.
2058 (pass_manager::register_dump_files): Adjust.
2060 2014-03-20 Mark Wielaard <mjw@redhat.com>
2062 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
2063 then represent the bound as normal constant value.
2065 2014-04-17 Jakub Jelinek <jakub@redhat.com>
2068 Forward port from 4.8 branch
2069 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
2071 * config/i386/bmiintrin.h (_blsi_u32): New.
2075 (_blsmsk_u32): Ditto.
2076 (_blsmsk_u64): Ditto.
2077 (_tzcnt_u32): Ditto.
2078 (_tzcnt_u64): Ditto.
2080 2014-04-17 Kito Cheng <kito@0xlab.org>
2082 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
2084 2014-04-17 Richard Biener <rguenther@suse.de>
2087 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
2088 boolean results for comparisons.
2090 2014-04-17 Richard Biener <rguenther@suse.de>
2092 PR tree-optimization/60836
2093 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
2094 initial PHI args to be gimple values.
2096 2014-04-17 Richard Biener <rguenther@suse.de>
2098 PR tree-optimization/60841
2099 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
2100 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
2101 of stmts to SLP build.
2102 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
2103 (vect_analyze_slp): Likewise.
2104 (vect_analyze_slp_instance): Likewise.
2105 (vect_build_slp_tree): Limit overall SLP tree growth.
2106 * tree-vectorizer.h (vect_analyze_data_refs,
2107 vect_analyze_slp): Adjust prototypes.
2109 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2111 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
2114 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2116 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
2117 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
2118 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
2119 for TARGET_SLOW_PSHUFB
2121 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2123 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
2124 * config/i386/i386.c (intel_cost): Ditto.
2126 2014-04-17 Joey Ye <joey.ye@arm.com>
2128 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
2130 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2132 * opts.c (common_handle_option): Disable -fipa-reference coorectly
2135 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2137 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
2138 (type_all_derivations_known_p): New predicate.
2139 (type_all_ctors_visible_p): New predicate.
2140 (type_possibly_instantiated_p): New predicate.
2141 (get_odr_type): Compute all_derivations_known.
2142 (dump_odr_type): Dump the flag.
2143 (maybe_record_type): Cleanup.
2144 (record_target_from_binfo): Add bases_to_consider array;
2145 record bases for types w/o instances and skip CXX destructor.
2146 (possible_polymorphic_call_targets_1): Add bases_to_consider
2147 and consider_construction parameters; check if type may have instance.
2148 (get_polymorphic_call_info): Set maybe_in_construction to true
2149 when we know nothing.
2150 (record_targets_from_bases): Skip CXX destructors; they are
2151 never called for types in construction.
2152 (possible_polymorphic_call_targets): Do not record target when
2153 type may not have instance.
2155 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2158 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
2159 external aliases alive, too.
2161 2014-04-16 Andrew Pinski <apinski@cavium.com>
2163 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
2166 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2168 * final.c (compute_alignments): Do not apply loop alignment to a block
2169 falling through to the exit.
2171 2014-04-16 Catherine Moore <clm@codesourcery.com>
2173 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
2174 Adjust constraints for microMIPS store patterns.
2176 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2178 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
2180 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2182 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
2183 (append_use): Run at -O0.
2184 (append_vdef): Likewise.
2185 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
2186 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
2188 2014-04-16 Jakub Jelinek <jakub@redhat.com>
2190 PR tree-optimization/60844
2191 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
2192 (propagate_op_to_single_use, remove_visited_stmt_chain,
2193 linearize_expr, repropagate_negates, reassociate_bb): Use it
2194 instead of gsi_remove.
2196 2014-04-16 Martin Jambor <mjambor@suse.cz>
2198 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
2199 ipa_transforms_to_apply.
2200 (cgraph_function_versioning): Assert that old_node has empty
2201 ipa_transforms_to_apply.
2202 * trans-mem.c (ipa_tm_create_version): Likewise.
2203 * tree-inline.c (tree_function_versioning): Do not duplicate
2204 ipa_transforms_to_apply.
2206 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2209 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
2211 Pass necessary as flags on 64-bit Solaris/x86.
2212 Use lowercase relocs for x86_64-*-*.
2213 * configure: Regenerate.
2215 2014-04-15 Jan Hubicka <jh@suse.cz>
2217 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
2218 (maybe_record_node, likely_target_p): Use it.
2220 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2223 Revert following patch
2225 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2228 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2229 software floating point or no floating point registers, do not
2230 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2231 in GPRs that occurs after we tested for GPRs that would never be
2234 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2235 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2236 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2237 specifically allow DDmode, since that does not use the SPE SIMD
2240 2014-03-21 Mark Wielaard <mjw@redhat.com>
2242 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
2243 as unsigned or int depending on type and value used.
2245 2014-04-15 Richard Biener <rguenther@suse.de>
2247 PR rtl-optimization/56965
2248 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
2249 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
2251 * alias.c (true_dependence_1): Do not call
2252 nonoverlapping_component_refs_p.
2253 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
2254 nonoverlapping_component_refs_p.
2255 (indirect_refs_may_alias_p): Likewise.
2257 2014-04-15 Teresa Johnson <tejohnson@google.com>
2259 * cfg.c (dump_bb_info): Fix flags check.
2260 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
2262 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2264 PR rtl-optimization/60663
2265 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2268 2014-04-15 Richard Biener <rguenther@suse.de>
2270 * lto-streamer.h (LTO_major_version): Bump to 4.
2272 2014-04-15 Richard Biener <rguenther@suse.de>
2274 * common.opt (lto_partition_model): New enum.
2275 (flto-partition=): Merge separate options with a single with argument,
2276 add -flto-partition=one support.
2277 * flag-types.h (enum lto_partition_model): Declare.
2278 * opts.c (finish_options): Remove duplicate -flto-partition=
2280 * lto-wrapper.c (run_gcc): Adjust.
2282 2014-04-15 Richard Biener <rguenther@suse.de>
2284 * alias.c (ncr_compar): New function.
2285 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2287 2014-04-15 Richard Biener <rguenther@suse.de>
2289 * alias.c (record_component_aliases): Do not walk BINFOs.
2291 2014-04-15 Richard Biener <rguenther@suse.de>
2293 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2294 Add struct function argument and adjust.
2295 (find_func_aliases_for_call): Likewise.
2296 (find_func_aliases): Likewise.
2297 (find_func_clobbers): Likewise.
2298 (intra_create_variable_infos): Likewise.
2299 (compute_points_to_sets): Likewise.
2300 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2302 2014-04-15 Richard Biener <rguenther@suse.de>
2304 * tree.c (iterative_hash_expr): Use enum tree_code_class
2305 to store TREE_CODE_CLASS.
2306 (tree_block): Likewise.
2307 (tree_set_block): Likewise.
2308 * tree.h (fold_build_pointer_plus_loc): Use
2309 convert_to_ptrofftype_loc.
2311 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2314 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2317 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2319 * cfgloop.h (struct loop): Move force_vectorize down.
2320 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2321 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2322 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2323 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2324 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2325 * tree-core.h (enum annot_expr_kind): Add new kind values.
2326 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2327 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2329 * tree.def (ANNOTATE_EXPR): Tweak comment.
2331 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2333 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2336 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2338 * tree.h (TYPE_IDENTIFIER): Declare.
2339 * tree.c (subrange_type_for_debug_p): Use it.
2340 * godump.c (go_format_type): Likewise.
2341 * dwarf2out.c (is_cxx_auto, modified_type_die,
2342 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2343 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2345 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2348 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2350 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2352 * config/i386/i386.c (examine_argument): Return bool. Return true if
2353 parameter should be passed in memory.
2354 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2355 (construct_container): Update calls to examine_argument.
2356 (function_arg_advance_64): Ditto.
2357 (return_in_memory_32): Merge with ix86_return_in_memory.
2358 (return_in_memory_64): Ditto.
2359 (return_in_memory_ms_64): Ditto.
2361 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2363 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2364 * coverage.c (coverage_compute_profile_id): Handle externally visible
2367 2014-04-14 Martin Jambor <mjambor@suse.cz>
2369 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2370 DECL_DISREGARD_INLINE_LIMITS functions.
2372 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2375 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2377 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2380 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2381 optimize_insn_for_speed_p instead of
2382 optimize_function_for_speed_p.
2384 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2386 * doc/invoke.texi (free): Document AArch64.
2388 2014-04-14 Richard Biener <rguenther@suse.de>
2390 PR tree-optimization/60042
2391 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2392 (insert_into_preds_of_block): Do not prevent PHI insertion
2393 for REFERENCE exprs here ...
2394 (eliminate_dom_walker::before_dom_children): ... but prevent
2395 their use here under similar conditions when applied to the
2396 IL after PRE optimizations.
2398 2014-04-14 Richard Biener <rguenther@suse.de>
2400 * passes.def: Move early points-to after early SRA.
2402 2014-04-14 Richard Biener <rguenther@suse.de>
2404 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2405 check for which sign-changes we allow when forwarding
2406 a converted value into a switch.
2408 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2410 * stor-layout.c (place_field): Finalize non-constant offset for the
2413 2014-04-14 Richard Biener <rguenther@suse.de>
2415 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2417 (expand_switch_using_bit_tests_p): Likewise.
2418 (process_switch): Compute and pass on speed_p based on the
2420 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2421 optimize_bb_for_speed_p.
2423 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2425 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2426 * function.h (struct function): Rename has_force_vect_loops into
2427 has_force_vectorize_loops.
2428 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2429 (input_struct_function_base): Likewise.
2430 * lto-streamer-out.c (output_cfg): Likewise.
2431 (output_struct_function_base): Likewise.
2432 * omp-low.c (expand_omp_simd): Likewise.
2433 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2434 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2435 (version_loop_for_if_conversion): Likewise.
2436 (tree_if_conversion): Likewise.
2437 (main_tree_if_conversion): Likewise.
2438 (gate_tree_if_conversion): Likewise.
2439 * tree-inline.c (copy_loops): Likewise.
2440 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2441 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2442 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2443 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2444 * tree-vectorizer.c (vectorize_loops): Likewise.
2445 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2447 2014-04-14 Richard Biener <rguenther@suse.de>
2450 * lto-streamer-out.c (wrap_refs): New function.
2451 (lto_output): Wrap symbol references in global initializes in
2452 type-preserving MEM_REFs.
2454 2014-04-14 Christian Bruel <christian.bruel@st.com>
2456 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2458 2014-04-14 Christian Bruel <christian.bruel@st.com>
2460 * config/sh/sh.md (setmemqi): New expand pattern.
2461 * config/sh/sh.h (CLEAR_RATIO): Define.
2462 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2463 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2465 2014-04-14 Richard Biener <rguenther@suse.de>
2468 * fold-const.c (negate_expr_p): Don't negate directional rounding
2470 (fold_negate_expr): Likewise.
2472 2014-04-14 Richard Biener <rguenther@suse.de>
2474 PR tree-optimization/59817
2475 PR tree-optimization/60453
2476 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2477 recursion to catch all CHRECs in the scalar evolution and restrict
2478 the predicate for the remains appropriately.
2480 2014-04-12 Catherine Moore <clm@codesourcery.com>
2482 * config/mips/constraints.md: Add new register constraint "kb".
2483 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2484 (*movhi_internal): Likewise.
2485 (*movqi_internal): Likewise.
2486 * config/mips/mips.h (M16_STORE_REGS): New register class.
2487 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2488 (REG_CLASS_CONTENTS): Likewise.
2489 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2491 2014-04-11 Tobias Burnus <burnus@net-b.de>
2494 * doc/invoke.texi (-Wformat-signedness): Document it.
2495 (Wformat=2): Mention that this enables -Wformat-signedness.
2497 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2499 * common/config/epiphany/epiphany-common.c
2500 (epiphany_option_optimization_table): Enable section anchors by
2501 default at -O1 or higher.
2502 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2503 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2504 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2505 carries no extra cost.
2506 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2507 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2508 * config/epiphany/predicates.md (memclob_operand): New predicate.
2509 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2510 Use memclob_operand predicate and X constraint for operand 3.
2512 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2514 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2515 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2518 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2520 PR rtl-optimization/60651
2521 * mode-switching.c (optimize_mode_switching): Make sure to emit
2522 sets of a lower numbered entity before sets of a higher numbered
2523 entity to a mode of the same or lower priority.
2524 When creating a seginfo for a basic block that starts with a code
2525 label, move the insertion point past the code label.
2526 (new_seginfo): Document and enforce requirement that
2527 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2528 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2529 * doc/tm.texi: Regenerate.
2531 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2534 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2536 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2538 * BASE-VER: Set to 4.10.0.
2540 2014-04-11 Tobias Burnus <burnus@net-b.de>
2543 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2544 * doc/gcc.texi (Service): Update description in the @menu
2545 * doc/invoke.texi (Option Summary): Remove misplaced and
2548 2014-04-11 Steve Ellcey <sellcey@mips.com>
2549 Jakub Jelinek <jakub@redhat.com>
2552 * expr.c (convert_move): Use emit_store_flag_force instead of
2553 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2556 2014-04-11 Richard Biener <rguenther@suse.de>
2559 * varasm.c (assemble_alias): Avoid endless error reporting
2560 recursion by setting TREE_ASM_WRITTEN.
2562 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2564 * config/s390/s390.md: Add a splitter for NOT rtx.
2566 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2568 PR rtl-optimization/60663
2569 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2571 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2572 Jakub Jelinek <jakub@redhat.com>
2575 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2576 flag from decl_node to node.
2578 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2581 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2582 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2583 ameliorating the cases where it can be.
2585 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2588 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2590 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2591 (loadsync_<mode>): Change mode.
2592 (load_quadpti, store_quadpti): New.
2593 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2594 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2595 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2597 2014-04-09 Cong Hou <congh@google.com>
2600 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2603 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2605 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2606 instead of vnor to exploit possible fusion opportunity in the
2608 (altivec_expand_vec_perm_const_le): Likewise.
2610 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2612 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2613 (loadsync_<mode>): Change mode.
2614 (load_quadpti, store_quadpti): New.
2615 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2616 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2618 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2621 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2622 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2623 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2625 2014-04-08 Richard Biener <rguenther@suse.de>
2628 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2629 a 64bit widest int print double-int similar to on HWI64 hosts.
2631 2014-04-08 Richard Biener <rguenther@suse.de>
2633 PR tree-optimization/60785
2634 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2635 default defs properly.
2637 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2639 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2640 (Weffc++): Likewise.
2642 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2644 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2645 set completep to false rather than true.
2647 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2650 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2651 ARM_TARGET2_DWARF_FORMAT.
2653 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2656 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2657 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2660 2014-04-07 Richard Biener <rguenther@suse.de>
2662 PR tree-optimization/60766
2663 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2664 (may_eliminate_iv): Convert cand_value_at result to desired type.
2666 2014-04-07 Jason Merrill <jason@redhat.com>
2669 * common.opt (-fno-gnu-unique): Add.
2670 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2672 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2674 * haifa-sched.c: Fix outdated function reference and minor
2675 grammar errors in introductory comment.
2677 2014-04-07 Richard Biener <rguenther@suse.de>
2680 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2682 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2684 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2687 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2688 size accounting for thunks.
2689 (pa_asm_output_mi_thunk): Use final_start_function() and
2690 final_end_function() to output function start and end directives.
2692 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2694 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2695 device specific ISA/ feature information. Remove short_sp and
2696 errata_skip ds. Add avr_device_specific_features enum to have device
2698 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2699 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2700 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2701 updated device specific info.
2702 * config/avr/avr-mcus.def: Merge device specific details to
2703 dev_attribute field.
2704 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2706 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2707 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2708 assembler if RMW isa supported by current device.
2709 * config/avr/genmultilib.awk: Update as device info structure changed.
2710 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2712 2014-04-04 Cong Hou <congh@google.com>
2714 PR tree-optimization/60656
2715 * tree-vect-stmts.c (supportable_widening_operation):
2716 Fix a bug that elements in a vector with vect_used_by_reduction
2717 property are incorrectly reordered when the operation on it is not
2718 consistant with the one in reduction operation.
2720 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2722 PR rtl-optimization/60155
2723 * gcse.c (record_set_data): New function.
2724 (single_set_gcse): New function.
2725 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2726 (hoist_code): Likewise.
2727 (get_pressure_class_and_nregs): Likewise.
2729 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2731 * explow.c (probe_stack_range): Emit a final optimization blockage.
2733 2014-04-04 Anthony Green <green@moxielogic.com>
2735 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2738 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2741 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2742 flags are set only during streaming.
2743 * ipa.c (process_references, walk_polymorphic_call_targets,
2744 symtab_remove_unreachable_nodes): Drop bodies of always inline
2745 after early inlining.
2746 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2748 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2749 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2752 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2755 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2758 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2760 (cortex_a53_fdivd): Likewise.
2762 2014-04-04 Martin Jambor <mjambor@suse.cz>
2765 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2767 * cgraph.c (clone_of_p): Also return true if thunks match.
2768 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2769 cgraph_function_or_thunk_node and an obsolete comment.
2770 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2772 (build_function_decl_skip_args): Likewise.
2773 (set_new_clone_decl_and_node_flags): New function.
2774 (duplicate_thunk_for_node): Likewise.
2775 (redirect_edge_duplicating_thunks): Likewise.
2776 (cgraph_clone_node): New parameter args_to_skip, pass it to
2777 redirect_edge_duplicating_thunks which is called instead of
2778 cgraph_redirect_edge_callee.
2779 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2780 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2782 2014-04-04 Jeff Law <law@redhat.com>
2785 * config/arm/predicates.md (const_int_I_operand): New predicate.
2786 (const_int_M_operand): Similarly.
2787 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2789 (insv_t2, extv_reg, extzv_t2): Likewise.
2790 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2791 (pop_multiple_with_writeback_and_return): Likewise.
2792 (vfp_pop_multiple_with_writeback): Likewise
2794 2014-04-04 Richard Biener <rguenther@suse.de>
2797 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2798 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2800 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2801 * gimplify.c (gimple_add_tmp_var_fn): New function.
2802 * gimple-expr.h (create_tmp_reg_fn): Declare.
2803 * gimple-expr.c (create_tmp_reg_fn): New function.
2804 * gimple-low.c (record_vars_into): Don't change cfun.
2805 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2806 code generation without cfun.
2808 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2811 * Makefile.in (install-driver): Fix shell scripting.
2813 2014-04-03 Cong Hou <congh@google.com>
2815 PR tree-optimization/60505
2816 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2817 threshold of number of iterations below which no vectorization
2819 * tree-vect-loop.c (new_loop_vec_info):
2820 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2821 * tree-vect-loop.c (vect_analyze_loop_operations):
2822 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2823 * tree-vect-loop.c (vect_transform_loop):
2824 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2825 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2826 of iterations of the loop and see if we should build the epilogue.
2828 2014-04-03 Richard Biener <rguenther@suse.de>
2830 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2831 (streamer_tree_cache_create): Adjust.
2832 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2833 to allow optional nodes array.
2834 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2835 (streamer_tree_cache_append): Likewise.
2836 (streamer_tree_cache_create): Create nodes array optionally
2837 as specified by parameter.
2838 * lto-streamer-out.c (create_output_block): Avoid maintaining
2839 the node array in the writer cache.
2840 (DFS_write_tree): Remove assertion.
2841 (produce_asm_for_decls): Free the out decl state hash table early.
2842 * lto-streamer-in.c (lto_data_in_create): Adjust for
2843 streamer_tree_cache_create prototype change.
2845 2014-04-03 Richard Biener <rguenther@suse.de>
2847 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2848 set TREE_CHAIN to NULL_TREE.
2850 2014-04-03 Richard Biener <rguenther@suse.de>
2852 PR tree-optimization/60740
2853 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2854 over all GIMPLE_COND operands.
2856 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2858 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2859 (Weffc++): Remove Scott's numbering, merge lists and reference
2862 2014-04-03 Nick Clifton <nickc@redhat.com>
2864 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2867 2014-04-03 Martin Jambor <mjambor@suse.cz>
2869 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2870 mention gcc_unreachable before failing.
2871 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2874 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2877 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2878 inconsistent code and instead mark the context inconsistent.
2879 (possible_polymorphic_call_targets): For inconsistent contexts
2880 return empty complete list.
2882 2014-04-02 Anthony Green <green@moxielogic.com>
2884 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2885 (extendqisi2, extendhisi2): Define.
2886 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2887 (WCHAR_TYPE): Change to unsigned int.
2889 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2891 PR tree-optimization/60733
2892 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2893 insertion point for PHI candidates to be the end of the feeding
2894 block for the PHI argument.
2896 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2898 PR rtl-optimization/60650
2899 * lra-constraints.c (process_alt_operands): Decrease reject for
2900 earlyclobber matching.
2902 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2904 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2906 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2908 * config/spu/spu.c (pad_bb): Do not crash when the last
2909 insn is CODE_FOR_blockage.
2911 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2913 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2914 lies outside the target mode.
2916 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2919 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2920 software floating point or no floating point registers, do not
2921 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2922 in GPRs that occurs after we tested for GPRs that would never be
2925 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2926 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2927 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2928 specifically allow DDmode, since that does not use the SPE SIMD
2931 2014-04-02 Richard Biener <rguenther@suse.de>
2934 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2935 MODE_INTs. Properly use negv_optab.
2936 (expand_abs): Likewise.
2938 2014-04-02 Richard Biener <rguenther@suse.de>
2941 * Makefile.in (install-driver): Guard extra installs with special
2944 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2946 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2949 2014-04-01 Richard Henderson <rth@redhat.com>
2952 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2953 alternative enabled before register allocation.
2955 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
2957 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
2958 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
2960 (nios2_large_got_address): Remove unneeded 'sym' parameter.
2961 (nios2_got_address): Update nios2_large_got_address call site.
2962 (nios2_delegitimize_address): New function.
2963 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
2964 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
2965 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
2967 2014-04-01 Martin Husemann <martin@duskware.de>
2969 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
2972 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
2974 PR rtl-optimization/60604
2975 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
2976 check from register_operand.
2977 (register_operand): Redefine in terms of general_operand.
2978 (nonmemory_operand): Use register_operand for the non-constant cases.
2980 2014-04-01 Richard Biener <rguenther@suse.de>
2982 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
2984 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
2986 * doc/invoke.texi (mapp-regs): Clarify.
2988 2014-03-31 Ulrich Drepper <drepper@gmail.com>
2990 * config/i386/avx512fintrin.h (__v32hi): Define type.
2991 (__v64qi): Likewise.
2992 (_mm512_set1_epi8): Define.
2993 (_mm512_set1_epi16): Define.
2994 (_mm512_set4_epi32): Define.
2995 (_mm512_set4_epi64): Define.
2996 (_mm512_set4_pd): Define.
2997 (_mm512_set4_ps): Define.
2998 (_mm512_setr4_epi64): Define.
2999 (_mm512_setr4_epi32): Define.
3000 (_mm512_setr4_pd): Define.
3001 (_mm512_setr4_ps): Define.
3002 (_mm512_setzero_epi32): Define.
3004 2014-03-31 Martin Jambor <mjambor@suse.cz>
3007 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
3008 callsite_arguments_match_p. Updated all callers. Also check types of
3009 corresponding formal parameters and actual arguments.
3010 (not_all_callers_have_enough_arguments_p) Renamed to
3011 some_callers_have_mismatched_arguments_p.
3013 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
3015 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
3017 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
3020 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
3023 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
3025 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
3026 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
3028 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
3029 Use FMAMODE_NOVF512 mode iterator.
3030 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
3031 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
3032 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
3033 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
3035 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
3036 Use VF_128_256 mode iterator.
3037 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
3040 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3042 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
3043 static chain if needed.
3045 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3048 * lra-constraints.c (index_part_to_reg): New.
3049 (process_address): Use it.
3051 2014-03-27 Jeff Law <law@redhat.com>
3052 Jakub Jelinek <jakub@redhat.com>
3055 * expr.c (do_tablejump): Use simplify_gen_binary rather than
3056 gen_rtx_{PLUS,MULT} to build up the address expression.
3058 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
3059 creating non-canonical RTL.
3061 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3064 * ipa-inline.c (want_inline_small_function_p): Short circuit large
3065 functions; reorganize to make cheap checks first.
3066 (inline_small_functions): Do not estimate growth when dumping;
3068 * ipa-inline.h (inline_summary): Add min_size.
3069 (growth_likely_positive): New function.
3070 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
3071 (set_cond_stmt_execution_predicate): Cleanup.
3072 (estimate_edge_size_and_time): Compute min_size.
3073 (estimate_calls_size_and_time): Likewise.
3074 (estimate_node_size_and_time): Likewise.
3075 (inline_update_overall_summary): Update min_size.
3076 (do_estimate_edge_time): Likewise.
3077 (do_estimate_edge_size): Update.
3078 (do_estimate_edge_hints): Update.
3079 (growth_likely_positive): New function.
3081 2014-03-28 Jakub Jelinek <jakub@redhat.com>
3084 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
3085 also if addr has VOIDmode.
3087 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3089 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
3090 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
3092 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
3093 instructions as well as AdvancedSIMD loads.
3095 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3097 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
3098 Use crypto_aese type.
3099 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
3100 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
3101 crypto_aese, crypto_aesmc. Move to types.md.
3102 * config/arm/types.md (crypto_aes): Split into crypto_aese,
3104 * config/arm/iterators.md (crypto_type): Likewise.
3106 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3108 * cgraph.c: Include expr.h and tree-dfa.h.
3109 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
3112 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3115 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
3116 regs from checking multi-reg pseudos.
3118 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3120 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
3122 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3124 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
3125 if it would clobber the stack pointer, even temporarily.
3127 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
3129 * mode-switching.c: Make small adjustments to the top comment.
3131 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3133 * config/rs6000/constraints.md (wD constraint): New constraint to
3134 match the constant integer to get the top DImode/DFmode out of a
3135 vector in a VSX register.
3137 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
3138 match the constant integer to get the top DImode/DFmode out of a
3139 vector in a VSX register.
3141 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
3144 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
3147 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
3148 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
3150 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
3151 Optimize vec_extract of 64-bit values, where the value being
3152 extracted is in the top word, where we can use scalar
3153 instructions. Add direct move and store support. Combine the big
3154 endian/little endian vector select load support into a single insn.
3155 (vsx_extract_<mode>_internal1): Likewise.
3156 (vsx_extract_<mode>_internal2): Likewise.
3157 (vsx_extract_<mode>_load): Likewise.
3158 (vsx_extract_<mode>_store): Likewise.
3159 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
3160 combined into vsx_extract_<mode>_load.
3161 (vsx_extract_<mode>_one_le): Likewise.
3163 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
3164 define the top 64-bit vector element.
3166 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
3169 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3170 Document vec_vbpermq builtin.
3173 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
3174 enable use of xxsldwi and xxpermdi builtin functions.
3175 (vec_xxpermdi): Likewise.
3177 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3178 Document use of vec_xxsldwi and vec_xxpermdi builtins.
3180 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
3182 PR rtl-optimization/60650
3183 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
3185 (find_spills_for): New.
3186 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
3187 Spill all pseudos on the second iteration.
3189 2014-03-27 Marek Polacek <polacek@redhat.com>
3192 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
3195 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3197 * config/s390/s390.c (s390_can_use_return_insn): Check for
3198 call-saved FPRs on 31 bit.
3200 2014-03-27 Jakub Jelinek <jakub@redhat.com>
3203 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
3204 if they need regimplification, just drop them instead of
3205 calling gimple_regimplify_operands on them.
3207 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
3210 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
3211 (aarch64_frame_pointer_required): Adjust logic.
3212 (aarch64_can_eliminate): Adjust logic.
3213 (aarch64_override_options_after_change): Adjust logic.
3215 2014-03-27 Dehao Chen <dehao@google.com>
3217 * ipa-inline.c (early_inliner): Update node's inline info.
3219 2014-03-26 Dehao Chen <dehao@google.com>
3221 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
3222 compiler inserted conditional jumps for NAN float check.
3224 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3226 * ubsan.h (ubsan_create_data): Change second argument's type
3227 to const location_t *.
3228 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
3230 (ubsan_create_data): Change second argument to const location_t *PLOC.
3231 Create Loc field whenever PLOC is non-NULL.
3232 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
3233 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
3237 * real.c (real_to_integer2): Change type of low to UHWI.
3239 2014-03-26 Tobias Burnus <burnus@net-b.de>
3241 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
3242 (CILK_SELF_SPECS): New define.
3243 (driver_self_specs): Use it.
3245 2014-03-26 Richard Biener <rguenther@suse.de>
3247 * tree-pretty-print.c (percent_K_format): Implement special
3248 case for LTO and its stripped down BLOCK tree.
3250 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3253 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
3255 * tree-vrp.c (simplify_internal_call_using_ranges): If only
3256 one range is range_int_cst_p, but not both, at least optimize
3257 addition/subtraction of 0 and multiplication by 0 or 1.
3258 * gimple-fold.c (gimple_fold_call): Fold
3259 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
3260 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
3261 INTEGER_CSTs, try to fold at least x * 0 and y - y.
3263 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3265 PR rtl-optimization/60452
3266 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3267 <case REG>: Return 1 for invalid offsets from the frame pointer.
3269 2014-03-26 Marek Polacek <polacek@redhat.com>
3272 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3275 2014-03-26 Marek Polacek <polacek@redhat.com>
3278 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3281 2014-03-26 Marek Polacek <polacek@redhat.com>
3284 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3285 multiplication in unsigned type.
3287 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3289 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3291 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3293 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3295 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3298 * cif-code.def (UNREACHABLE) New code.
3299 * ipa-inline.c (inline_small_functions): Skip edges to
3300 __builtlin_unreachable.
3301 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3302 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3303 predicate to __bulitin_unreachable.
3304 (set_cond_stmt_execution_predicate): Fix issue when
3305 invert_tree_comparison returns ERROR_MARK.
3306 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3307 propagate to inline clones.
3308 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3310 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3311 * cgraphclones.c (cgraph_clone_node): If call destination is already
3312 ureachable, do not redirect it back.
3313 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3316 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3318 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3319 Do not modify inline clones.
3321 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3323 * config/i386/i386.md (general_sext_operand): New mode attr.
3324 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3325 don't generate (sign_extend (const_int)).
3326 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3327 operands[2]. Use We constraint instead of <i> and
3328 <general_sext_operand> predicate instead of <general_operand>.
3329 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3330 * config/i386/constraints.md (We): New constraint.
3331 * config/i386/predicates.md (x86_64_sext_operand,
3332 sext_operand): New predicates.
3334 2014-03-25 Martin Jambor <mjambor@suse.cz>
3337 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3338 inconsistent devirtualizations to __builtin_unreachable.
3340 2014-03-25 Marek Polacek <polacek@redhat.com>
3343 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3345 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3347 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3348 order of elements for big-endian.
3350 2014-03-25 Richard Biener <rguenther@suse.de>
3353 * gimplify-me.c (gimple_regimplify_operands): Update the
3356 2014-03-25 Martin Jambor <mjambor@suse.cz>
3359 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3360 (lto_output_varpool_node): Likewise.
3361 (input_overwrite_node): Likewise.
3362 (input_varpool_node): Likewise.
3364 2014-03-25 Richard Biener <rguenther@suse.de>
3366 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3367 (run_gcc): Likewise.
3369 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3371 * combine.c (simplify_compare_const): Add MODE argument.
3372 Handle mode_width 0 as very large mode_width.
3373 (try_combine, simplify_comparison): Adjust callers.
3375 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3376 type to avoid signed integer overflow.
3377 * explow.c (plus_constant): Likewise.
3379 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3381 * doc/generic.texi: Correct typos.
3383 2014-03-24 Tobias Burnus <burnus@net-b.de>
3385 * doc/invoke.texi (-flto): Expand section about
3386 using static libraries with LTO.
3388 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3390 PR rtl-optimization/60501
3391 * optabs.def (addptr3_optab): New optab.
3392 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3393 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3394 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3396 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3398 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3400 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3402 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3405 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3406 (_mm256_undefined_ps): Define.
3407 (_mm256_undefined_pd): Define.
3408 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3409 (_mm_undefined_pd): Define.
3410 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3411 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3412 (_mm512_undefined_ps): Define.
3413 (_mm512_undefined_pd): Define.
3414 Use _mm*_undefined_*.
3415 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3417 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3419 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3420 (lshr_simd): DI mode added.
3421 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3422 (aarch64_ushr_simddi): Likewise.
3423 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3424 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3425 (vshrd_n_u64): Likewise.
3427 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3429 * Makefile.in (s-macro_list): Depend on cc1.
3431 2014-03-23 Teresa Johnson <tejohnson@google.com>
3433 * ipa-utils.c (ipa_print_order): Use specified dump file.
3435 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3437 PR rtl-optimization/60601
3438 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3440 * gcc.c (eval_spec_function): Initialize save_growing_value.
3442 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3445 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3446 code == MINUS_EXPR, never swap op0 with op1.
3448 * toplev.c (init_local_tick): Avoid signed integer multiplication
3450 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3451 shift by first operand's bitsize.
3453 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3456 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3458 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3459 TARGET_ISA_64BIT_P(x).
3461 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3463 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3464 pattern for vector nor instead of subtract from splat(-1).
3465 (altivec_expand_vec_perm_const_le): Likewise.
3467 2014-03-21 Richard Henderson <rth@twiddle.net>
3470 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3471 related insns after epilogue_completed.
3473 2014-03-21 Martin Jambor <mjambor@suse.cz>
3476 * cgraph.h (symtab_node): New flag body_removed.
3477 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3478 when removing bodies.
3479 * symtab.c (dump_symtab_base): Dump body_removed flag.
3480 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3481 had their bodies removed.
3483 2014-03-21 Martin Jambor <mjambor@suse.cz>
3486 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3489 2014-03-21 Richard Biener <rguenther@suse.de>
3491 PR tree-optimization/60577
3492 * tree-core.h (struct tree_base): Document nothrow_flag use
3493 in VAR_DECL_NONALIASED.
3494 * tree.h (VAR_DECL_NONALIASED): New.
3495 (may_be_aliased): Adjust.
3496 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3498 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3500 * expr.c (expand_expr_real_1): Remove outdated comment.
3502 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3505 * ira.c (adjust_cleared_regs): Call copy_rtx on
3506 *reg_equiv[REGNO (loc)].src_p before passing it to
3507 simplify_replace_fn_rtx.
3510 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3511 into CONST, put pic register as first operand of PLUS. Use
3512 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3514 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3516 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3518 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3520 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3521 around for store forwarding issue in the FPU on the UT699.
3522 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3523 loads and operations if -mfix-ut699 is specified.
3524 (divtf3_hq): Tweak attribute.
3525 (sqrttf2_hq): Likewise.
3527 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3529 * calls.c (store_one_arg): Remove incorrect const qualification on the
3530 type of the temporary.
3531 * cfgexpand.c (expand_return): Likewise.
3532 * expr.c (expand_constructor): Likewise.
3533 (expand_expr_real_1): Likewise.
3535 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3537 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3540 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3543 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3545 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3547 * config/arm/aarch-common-protos.h
3548 (alu_cost_table): Fix spelling of "extend".
3549 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3551 2014-03-19 Richard Biener <rguenther@suse.de>
3554 * tree-core.h (tree_type_common): Re-order pointer members
3555 to reduce recursion depth during GC walks.
3557 2014-03-19 Marek Polacek <polacek@redhat.com>
3560 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3561 before accessing it.
3563 2014-03-19 Richard Biener <rguenther@suse.de>
3566 * lto-streamer-in.c (input_function): In WPA stage do not drop
3569 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3571 PR tree-optimization/60559
3572 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3573 with build_zero_cst assignment.
3575 2014-03-18 Kai Tietz <ktietz@redhat.com>
3577 PR rtl-optimization/56356
3578 * sdbout.c (sdbout_parms): Verify that parms'
3579 incoming argument is valid.
3580 (sdbout_reg_parms): Likewise.
3582 2014-03-18 Richard Henderson <rth@redhat.com>
3585 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3586 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3587 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3589 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3591 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3592 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3593 Italicize plugin event names in description. Explain that
3594 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3595 Remind that no GCC functions should be called after PLUGIN_FINISH.
3596 Explain what pragmas with expansion are.
3598 2014-03-18 Martin Liska <mliska@suse.cz>
3600 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3601 gimple call statement is update.
3602 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3603 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3605 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3608 * ubsan.c (ubsan_instrument_unreachable): Call
3609 initialize_sanitizer_builtins.
3610 (ubsan_pass): Likewise.
3613 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3614 varpool_finalize_decl instead of rest_of_decl_compilation.
3616 2014-03-18 Richard Biener <rguenther@suse.de>
3618 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3619 by using bitmap_and_compl instead of bitmap_and_compl_into.
3620 (df_rd_transfer_function): Likewise.
3622 2014-03-18 Richard Biener <rguenther@suse.de>
3624 * doc/lto.texi (fresolution): Fix typo.
3626 2014-03-18 Richard Biener <rguenther@suse.de>
3628 * doc/invoke.texi (flto): Update for changes in 4.9.
3630 2014-03-18 Richard Biener <rguenther@suse.de>
3632 * doc/loop.texi: Remove section on the removed lambda framework.
3633 Update loop docs with recent changes in preserving loop structure.
3635 2014-03-18 Richard Biener <rguenther@suse.de>
3637 * doc/lto.texi (-fresolution): Document.
3639 2014-03-18 Richard Biener <rguenther@suse.de>
3641 * doc/contrib.texi: Adjust my name.
3643 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3646 * internal-fn.c: Include diagnostic-core.h.
3647 (expand_BUILTIN_EXPECT): New function.
3648 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3649 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3650 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3651 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3653 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3654 Revert 3 argument __builtin_expect code.
3655 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3656 * gimple-fold.c (gimple_fold_call): Likewise.
3657 * tree.h (fold_builtin_expect): New prototype.
3658 * builtins.c (build_builtin_expect_predicate): Add predictor
3659 argument, if non-NULL, create 3 argument __builtin_expect.
3660 (fold_builtin_expect): No longer static. Add ARG2 argument,
3661 pass it through to build_builtin_expect_predicate.
3662 (fold_builtin_2): Adjust caller.
3663 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3664 * internal-fn.def (BUILTIN_EXPECT): New.
3666 2014-03-18 Tobias Burnus <burnus@net-b.de>
3669 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3670 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3671 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3673 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3676 * predict.c (combine_predictions_for_bb): Fix up formatting.
3677 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3678 fill what it points to if non-NULL.
3679 (tree_predict_by_opcode): Adjust caller, use the predictor.
3680 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3682 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3684 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3685 proper constant for the store mode.
3687 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3689 * symtab.c (change_decl_assembler_name): Fix transparent alias
3692 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3694 * config/aarch64/aarch64.c: Correct the comments about the
3695 aarch64 stack layout.
3697 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3699 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3700 check for GF_OMP_FOR_KIND_FOR.
3702 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3704 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3705 ymm and zmm register names.
3707 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3710 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3711 note creation for the 2010-08-31 changes.
3713 2014-03-17 Marek Polacek <polacek@redhat.com>
3716 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3717 as -fno-tree-loop-vectorize.
3718 (expand_omp_simd): Likewise.
3720 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3722 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3723 (eligible_for_call_delay): New prototype.
3724 * config/sparc/sparc.c (tls_call_delay): Rename into...
3725 (eligible_for_call_delay): ...this. Return false if the instruction
3726 cannot be put in the delay slot of a branch.
3727 (eligible_for_restore_insn): Simplify.
3728 (eligible_for_return_delay): Return false if the instruction cannot be
3729 put in the delay slot of a branch and simplify.
3730 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3731 put in the delay slot of a branch.
3732 * config/sparc/sparc.md (fix_ut699): New attribute.
3733 (tls_call_delay): Delete.
3734 (in_call_delay): Reimplement.
3735 (eligible_for_sibcall_delay): Rename into...
3736 (in_sibcall_delay): ...this.
3737 (eligible_for_return_delay): Rename into...
3738 (in_return_delay): ...this.
3739 (in_branch_delay): Reimplement.
3740 (in_uncond_branch_delay): Delete.
3741 (in_annul_branch_delay): Delete.
3743 2014-03-14 Richard Henderson <rth@redhat.com>
3746 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3747 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3748 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3749 (floathi splitters): Remove.
3750 (float<SWI48x>xf2): New pattern.
3751 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3752 code that tried to handle DImode for 32-bit, but which was excluded
3753 by the pattern's condition. Drop allocation of stack temporary.
3754 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3755 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3756 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3757 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3758 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3759 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3760 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3761 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3762 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3763 (*float<SWI48x><X87MODEF>2_i387): Remove.
3764 (all float _with_temp splitters): Remove.
3765 (*float<SWI48x><MODEF>2_i387): New pattern.
3766 (*float<SWI48><MODEF>2_sse): New pattern.
3767 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3768 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3770 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3771 Marek Polacek <polacek@redhat.com>
3774 * common.opt (dump_base_name_prefixed): New Variable.
3775 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3776 if x_dump_base_name_prefixed is already set, set it at the end.
3778 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3780 PR rtl-optimization/60508
3781 * lra-constraints.c (get_reload_reg): Add new parameter
3783 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3784 Pass the new parameter values.
3786 2014-03-14 Richard Biener <rguenther@suse.de>
3788 * common.opt: Revert unintented changes from r205065.
3791 2014-03-14 Richard Biener <rguenther@suse.de>
3794 * cfghooks.c (split_block): Properly adjust all loops the
3795 block was a latch of.
3797 2014-03-14 Martin Jambor <mjambor@suse.cz>
3800 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3803 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3806 * config/avr/avr.c (avr_set_current_function): Pass function name
3807 through default_strip_name_encoding before sanity checking instead
3808 of skipping the first char of the assembler name.
3810 2014-03-13 Richard Henderson <rth@redhat.com>
3813 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3814 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3815 * config/i386/i386-protos.h: Likewise.
3816 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3817 in the expander instead of a splitter.
3818 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3819 any possibility of requiring a memory.
3820 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3821 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3822 (fp branch splitters): Update for ix86_split_fp_branch.
3823 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3824 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3825 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3826 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3827 (*fop_<MODEF>_3_i387): Likewise.
3828 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3829 (splitters for the fop_* register patterns): Remove.
3830 (fscalexf4_i387): Rename from *fscalexf4_i387.
3831 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3833 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3835 PR tree-optimization/59779
3836 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3837 type for bitsize and maxsize instead of HOST_WIDE_INT.
3839 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3841 PR rtl-optimization/57320
3842 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3843 the CFG after thread_prologue_and_epilogue_insns.
3845 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3847 PR rtl-optimization/57189
3848 * lra-constraints.c (process_alt_operands): Disfavor spilling
3851 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3853 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3855 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3857 PR tree-optimization/59025
3859 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3860 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3862 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3865 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3866 calls of avr_out_plus_1.
3868 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3870 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3871 BB's single pred and update the father loop's latch info later.
3873 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3875 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3879 (VEC_base): Likewise.
3880 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3881 registers, we need to swap double words in little endian mode.
3883 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3884 to be a container mode for 128-bit integer operations added in ISA
3885 2.07. Unlike TImode and PTImode, the preferred register set is
3886 the Altivec/VMX registers for the 128-bit operations.
3888 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3890 (rs6000_split_128bit_ok_p): Likewise.
3892 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3893 macros for creating ISA 2.07 normal and overloaded builtin
3894 functions with 3 arguments.
3895 (BU_P8V_OVERLOAD_3): Likewise.
3896 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3897 for use as overloaded functions.
3898 (VPERM_1TI_UNS): Likewise.
3899 (VSEL_1TI): Likewise.
3900 (VSEL_1TI_UNS): Likewise.
3901 (ST_INTERNAL_1ti): Likewise.
3902 (LD_INTERNAL_1ti): Likewise.
3903 (XXSEL_1TI): Likewise.
3904 (XXSEL_1TI_UNS): Likewise.
3905 (VPERM_1TI): Likewise.
3906 (VPERM_1TI_UNS): Likewise.
3907 (XXPERMDI_1TI): Likewise.
3908 (SET_1TI): Likewise.
3909 (LXVD2X_V1TI): Likewise.
3910 (STXVD2X_V1TI): Likewise.
3911 (VEC_INIT_V1TI): Likewise.
3912 (VEC_SET_V1TI): Likewise.
3913 (VEC_EXT_V1TI): Likewise.
3914 (EQV_V1TI): Likewise.
3915 (NAND_V1TI): Likewise.
3916 (ORC_V1TI): Likewise.
3917 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3918 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3920 (VADDUQM): Likewise.
3921 (VSUBCUQ): Likewise.
3922 (VADDEUQM): Likewise.
3923 (VADDECUQ): Likewise.
3924 (VSUBEUQM): Likewise.
3925 (VSUBECUQ): Likewise.
3927 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3928 __int128_t and __uint128_t types.
3929 (__uint128_type): Likewise.
3930 (altivec_categorize_keyword): Add support for vector __int128_t,
3931 vector __uint128_t, vector __int128, and vector unsigned __int128
3932 as a container type for TImode operations that need to be done in
3933 VSX/Altivec registers.
3934 (rs6000_macro_to_expand): Likewise.
3935 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3936 to support 128-bit integer instructions vaddcuq, vadduqm,
3937 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3938 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3940 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
3941 for V1TImode, and set up preferences to use VSX/Altivec registers.
3942 Setup VSX reload handlers.
3943 (rs6000_debug_reg_global): Likewise.
3944 (rs6000_init_hard_regno_mode_ok): Likewise.
3945 (rs6000_preferred_simd_mode): Likewise.
3946 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
3947 (easy_altivec_constant): Likewise.
3948 (output_vec_const_move): Likewise.
3949 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3951 (rs6000_expand_vector_extract): Likewise.
3952 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3954 (rs6000_const_vec): Add support for V1TImode.
3955 (rs6000_emit_le_vsx_load): Swap double words when loading or
3956 storing TImode/V1TImode.
3957 (rs6000_emit_le_vsx_store): Likewise.
3958 (rs6000_emit_le_vsx_move): Likewise.
3959 (rs6000_emit_move): Add support for V1TImode.
3960 (altivec_expand_ld_builtin): Likewise.
3961 (altivec_expand_st_builtin): Likewise.
3962 (altivec_expand_vec_init_builtin): Likewise.
3963 (altivec_expand_builtin): Likewise.
3964 (rs6000_init_builtins): Add support for V1TImode type. Add
3965 support for ISA 2.07 128-bit integer builtins. Define type names
3966 for the VSX/Altivec vector types.
3967 (altivec_init_builtins): Add support for overloaded vector
3968 functions with V1TImode type.
3969 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
3970 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
3972 (rs6000_split_128bit_ok_p): Likewise.
3973 (rs6000_handle_altivec_attribute): Create V1TImode from vector
3974 __int128_t and vector __uint128_t.
3976 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
3977 and mode attributes.
3984 (VS_scalar): Likewise.
3985 (VS_double): Likewise.
3986 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
3988 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
3989 we support the ISA 2.07 128-bit integer arithmetic instructions.
3990 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
3991 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
3992 and TImode types for use with the builtin functions.
3993 (V1TI_type_node): Likewise.
3994 (unsigned_V1TI_type_node): Likewise.
3995 (intTI_type_internal_node): Likewise.
3996 (uintTI_type_internal_node): Likewise.
3998 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
3999 128-bit builtin functions.
4000 (UNSPEC_VADDEUQM): Likewise.
4001 (UNSPEC_VADDECUQ): Likewise.
4002 (UNSPEC_VSUBCUQ): Likewise.
4003 (UNSPEC_VSUBEUQM): Likewise.
4004 (UNSPEC_VSUBECUQ): Likewise.
4005 (VM): Add V1TImode to vector mode iterators.
4007 (VI_unit): Likewise.
4008 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
4009 (altivec_vaddcuq): Likewise.
4010 (altivec_vsubuqm): Likewise.
4011 (altivec_vsubcuq): Likewise.
4012 (altivec_vaddeuqm): Likewise.
4013 (altivec_vaddecuq): Likewise.
4014 (altivec_vsubeuqm): Likewise.
4015 (altivec_vsubecuq): Likewise.
4017 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
4019 (BOOL_128): Likewise.
4020 (BOOL_REGS_OUTPUT): Likewise.
4021 (BOOL_REGS_OP1): Likewise.
4022 (BOOL_REGS_OP2): Likewise.
4023 (BOOL_REGS_UNARY): Likewise.
4024 (BOOL_REGS_AND_CR0): Likewise.
4026 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
4027 128-bit integer builtin support.
4028 (vec_vadduqm): Likewise.
4029 (vec_vaddecuq): Likewise.
4030 (vec_vaddeuqm): Likewise.
4031 (vec_vsubecuq): Likewise.
4032 (vec_vsubeuqm): Likewise.
4033 (vec_vsubcuq): Likewise.
4034 (vec_vsubuqm): Likewise.
4036 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4037 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
4038 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
4039 128-bit integer add/subtract to ISA 2.07.
4041 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
4043 * config/arc/arc.c (arc_predicate_delay_insns):
4044 Fix third argument passed to conditionalize_nonjump.
4046 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
4048 * config/aarch64/aarch64-builtins.c
4049 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
4050 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
4051 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
4052 instead of __builtin_lfloor.
4053 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
4055 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4057 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
4058 (tree_ssa_ifcombine_bb_1): New function.
4059 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
4060 is an empty forwarder block to then_bb or vice versa and then_bb
4061 and else_bb are effectively swapped.
4063 2014-03-12 Christian Bruel <christian.bruel@st.com>
4066 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
4067 REG_CFA_DEF_CFA note.
4068 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
4069 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
4071 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
4073 PR tree-optimization/60454
4074 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
4076 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4078 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
4079 Do not define target_cpu_default2 to generic.
4080 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
4081 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
4082 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
4084 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4085 Marc Glisse <marc.glisse@inria.fr>
4087 PR tree-optimization/60502
4088 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
4089 instead of build_low_bits_mask.
4091 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4094 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
4095 if there are multiple uses, but op doesn't live on E edge.
4096 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
4097 clobber stmts before __builtin_unreachable.
4099 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
4101 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
4102 hard_frame_pointer_rtx.
4103 * cse.c (cse_insn): Remove volatile check.
4104 * cselib.c (cselib_process_insn): Likewise.
4105 * dse.c (scan_insn): Likewise.
4107 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4109 * config/arc/arc.c (conditionalize_nonjump): New function,
4111 (arc_ifcvt): ... this.
4112 (arc_predicate_delay_insns): Use it.
4114 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4116 * config/arc/predicates.md (extend_operand): During/after reload,
4117 allow const_int_operand.
4118 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
4119 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
4120 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
4122 (umulsi3_highpart_i): Likewise.
4124 2014-03-11 Richard Biener <rguenther@suse.de>
4126 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
4127 Add asserts to guard possible wrong-code bugs.
4129 2014-03-11 Richard Biener <rguenther@suse.de>
4131 PR tree-optimization/60429
4132 PR tree-optimization/60485
4133 * tree-ssa-structalias.c (set_union_with_increment): Properly
4134 take into account all fields that overlap the shifted vars.
4135 (do_sd_constraint): Likewise.
4136 (do_ds_constraint): Likewise.
4137 (get_constraint_for_ptr_offset): Likewise.
4139 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
4141 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
4142 (nios2_compute_frame_layout):
4143 Add calculation of cfun->machine->fp_save_offset.
4144 (nios2_expand_prologue): Correct setting of frame pointer register
4146 (nios2_expand_epilogue): Update recovery of stack pointer from
4147 frame pointer accordingly.
4148 (nios2_initial_elimination_offset): Update calculation of offset
4149 for eliminating to HARD_FRAME_POINTER_REGNUM.
4151 2014-03-10 Jakub Jelinek <jakub@redhat.com>
4154 * ipa.c (symtab_remove_unreachable_nodes): Don't call
4155 cgraph_get_create_node on VAR_DECLs.
4157 2014-03-10 Richard Biener <rguenther@suse.de>
4160 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
4162 2014-03-08 Douglas B Rupp <rupp@gnat.com>
4164 * config/vms/vms.opt (vms_float_format): New variable.
4166 2014-03-08 Tobias Burnus <burnus@net-b.de>
4168 * doc/invoke.texi (-fcilkplus): Update implementation status.
4170 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
4171 Richard Biener <rguenther@suse.de>
4173 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
4174 consistently accross all TUs.
4175 (run_gcc): Enable -fshort-double automatically at link at link-time
4176 and disallow override.
4178 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
4181 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
4182 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
4183 if they can't be used.
4185 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4187 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
4188 for Solaris 11/x86 ld.
4189 * configure: Regenerate.
4191 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4193 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
4194 (LIB_TLS_SPEC): Save as ld_tls_libs.
4195 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
4196 (HAVE_AS_IX86_TLSLDM): New test.
4197 * configure, config.in: Regenerate.
4198 * config/i386/i386.c (legitimize_tls_address): Fall back to
4199 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
4200 cannot support TLS_MODEL_LOCAL_DYNAMIC.
4201 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
4202 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
4204 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
4206 * common.opt (fira-loop-pressure): Mark as optimization.
4208 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
4210 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
4211 an OpenMP mappable type.
4213 2014-03-06 Matthias Klose <doko@ubuntu.com>
4215 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
4216 MULTILIB_OSDIRNAMES is not defined.
4218 2014-03-06 Jakub Jelinek <jakub@redhat.com>
4219 Meador Inge <meadori@codesourcery.com>
4222 * config/arm/arm.c (arm_tls_symbol_p): Remove.
4223 (arm_legitimize_address): Call legitimize_tls_address for any
4224 arm_tls_referenced_p expression, handle constant addend. Call it
4225 before testing for !TARGET_ARM.
4226 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
4228 2014-03-06 Richard Biener <rguenther@suse.de>
4234 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4236 * tree-streamer.c (record_common_node): Assert we don't record
4237 nodes with type double.
4238 (preload_common_node): Skip type double, complex double and double
4239 pointer since it is now frontend dependent due to fshort-double option.
4241 2014-03-06 Richard Biener <rguenther@suse.de>
4243 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
4244 or -fno-lto is specified and the linker has full plugin support.
4245 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
4246 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
4247 * lto-wrapper.c (merge_and_complain): Merge compile-time
4248 optimization levels.
4249 (run_gcc): And pass it through to the link options.
4251 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
4255 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4257 * cselib.c (remove_useless_values): Skip to avoid quadratic
4258 behavior if the condition moved from...
4259 (cselib_process_insn): ... here holds.
4261 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4264 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4265 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4268 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4269 (TM_H): Add x86-tune.def.
4271 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4273 * config/aarch64/aarch64.c (generic_tunings):
4274 Use cortexa57_extra_costs.
4276 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4279 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4280 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4281 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4284 2014-03-04 Heiher <r@hev.cc>
4286 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4287 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4289 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4291 * config/i386/predicates.md (const2356_operand): Change to ...
4292 (const2367_operand): ... this.
4293 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4295 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4296 (*avx512pf_scatterpf<mode>sf): Ditto.
4297 (avx512pf_scatterpf<mode>df): Ditto.
4298 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4299 (*avx512pf_scatterpf<mode>df): Ditto.
4300 * config/i386/i386.c (ix86_expand_builtin): Update
4301 incorrect hint operand error message.
4303 2014-03-04 Richard Biener <rguenther@suse.de>
4305 * lto-section-in.c (lto_get_section_data): Fix const cast.
4307 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4309 * tree-streamer.c (record_common_node): Assert we don't record
4310 nodes with type double.
4311 (preload_common_node): Skip type double, complex double and double
4312 pointer since it is now frontend dependent due to fshort-double option.
4314 2014-03-04 Richard Biener <rguenther@suse.de>
4317 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4318 (lto_input_toplevel_asms): Likewise.
4319 * lto-section-in.c (lto_get_section_data): Instead do it here
4322 2014-03-04 Richard Biener <rguenther@suse.de>
4324 PR tree-optimization/60382
4325 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4326 dead PHIs a reduction.
4328 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4330 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4332 (_mm_prefetch): Move out of GCC target("sse") pragma.
4333 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4334 GCC target("prfchw") pragma.
4335 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4337 * config/i386/i386.c (ix86_option_override_internal): Enable
4338 -mprfchw with -mprefetchwt1.
4340 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4342 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4345 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4347 * opts.h (CL_PCH_IGNORE): Define.
4348 * targhooks.c (option_affects_pch_p):
4349 Return false for options that have CL_PCH_IGNORE set.
4350 * opt-functions.awk: Process PchIgnore.
4351 * doc/options.texi: Document PchIgnore.
4353 * config/arc/arc.opt (misize): Add PchIgnore property.
4355 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4357 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4358 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4359 constraint on constants to permit them being loaded into
4360 GENERAL_REGS or BASE_REGS.
4362 2014-03-03 Nick Clifton <nickc@redhat.com>
4364 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4365 anti-cacnonical alternatives.
4366 (negandhi3_real): New pattern.
4367 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4369 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4371 * config/avr/avr-mcus.def: Remove atxmega16x1.
4372 * config/avr/avr-tables.opt: Regenerate.
4373 * config/avr/t-multilib: Regenerate.
4374 * doc/avr-mmcu.texi: Regenerate.
4376 2014-03-03 Tobias Grosser <tobias@grosser.es>
4377 Mircea Namolaru <mircea.namolaru@inria.fr>
4379 PR tree-optimization/58028
4380 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4383 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4385 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4386 not handled by recognizers.
4388 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4391 * function.c (expand_function_end): Don't emit
4392 clobber_return_register sequence if clobber_after is a BARRIER.
4393 * cfgexpand.c (construct_exit_block): Append instructions before
4394 return_label to prev_bb.
4396 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4398 * config/rs6000/constraints.md: Document reserved use of "wc".
4400 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4403 * ipa.c (function_and_variable_visibility): When dissolving comdat
4404 group, also set all symbols to local.
4406 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4411 2013-12-14 Jan Hubicka <jh@suse.cz>
4413 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4415 2014-03-02 Jon Beniston <jon@beniston.com>
4421 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4422 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4423 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4424 (simple_return, *simple_return): New patterns
4425 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4426 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4428 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4430 * dwarf2out.c (gen_subprogram_die): Tidy.
4432 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4435 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4436 (*mov_t_msb_neg_negc): ... this new insn.
4438 2014-02-28 Jason Merrill <jason@redhat.com>
4441 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4444 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4447 * dwarf2out.c (decltype_auto_die): New static.
4448 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4449 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4450 (is_cxx_auto): Likewise.
4452 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4454 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4455 we are not using general regs only.
4457 2014-02-28 Richard Biener <rguenther@suse.de>
4460 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4461 previous fix and only allow to remove trivial pre-headers
4462 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4463 (remove_forwarder_block): Properly update the latch of a loop.
4465 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4468 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4469 (cselib_preserved_hash_table): New.
4470 (preserve_constants_and_equivs): Move preserved vals to it.
4471 (cselib_find_slot): Look it up first.
4472 (cselib_init): Initialize it.
4473 (cselib_finish): Release it.
4474 (dump_cselib_table): Dump it.
4476 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4479 * cselib.c (remove_useless_values): Skip to avoid quadratic
4480 behavior if the condition moved from...
4481 (cselib_process_insn): ... here holds.
4483 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4486 * var-tracking.c (vt_initialize): Apply the same condition to
4487 preserve the CFA base value.
4489 2014-02-28 Joey Ye <joey.ye@arm.com>
4492 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4493 if reload in progress or completed.
4495 2014-02-28 Tobias Burnus <burnus@net-b.de>
4498 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4501 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4503 * doc/tm.texi.in (Condition Code Status): Update documention for
4504 relative locations of cc0-setter and cc0-user.
4506 2014-02-27 Jeff Law <law@redhat.com>
4508 PR rtl-optimization/52714
4509 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4510 into two independent simple sets, if I3 is a jump, ensure the
4511 pattern we place into I3 is a (set (pc) ...).
4513 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4514 Jeff Law <law@redhat.com>
4516 PR rtl-optimization/49847
4517 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4518 are in different blocks.
4519 * doc/tm.texi (Condition Code Status): Update documention for
4520 relative locations of cc0-setter and cc0-user.
4522 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4525 * lra.c (lra_emit_add): Check SUBREG too.
4527 2014-02-27 Andreas Schwab <schwab@suse.de>
4529 * config/m68k/m68k.c (m68k_option_override): Disable
4530 -flive-range-shrinkage for classic m68k.
4531 (m68k_override_options_after_change): Likewise.
4533 2014-02-27 Marek Polacek <polacek@redhat.com>
4536 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4537 -Wmaybe-uninitialized.
4539 2014-02-27 Alan Modra <amodra@gmail.com>
4542 * reload1.c (emit_input_reload_insns): When reload_override_in,
4543 set old to rl->in_reg when rl->in_reg is a subreg.
4545 2014-02-26 Richard Biener <rguenther@suse.de>
4548 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4550 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4552 * common/config/i386/predicates.md (const1256_operand): Remove.
4553 (const2356_operand): New.
4554 (const_1_to_2_operand): Remove.
4555 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4556 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4557 (*avx512pf_gatherpf<mode>sf): Ditto.
4558 (avx512pf_gatherpf<mode>df): Ditto.
4559 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4560 (*avx512pf_gatherpf<mode>df): Ditto.
4561 (avx512pf_scatterpf<mode>sf): Ditto.
4562 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4563 (*avx512pf_scatterpf<mode>sf): Ditto.
4564 (avx512pf_scatterpf<mode>df): Ditto.
4565 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4566 (*avx512pf_scatterpf<mode>df): Ditto.
4567 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4569 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4571 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4572 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4573 (_mm512_mask_testn_epi64_mask): Move to ...
4574 * config/i386/avx512cdintrin.h: Here.
4575 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4576 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4577 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4578 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4579 TARGET_AVX512F from TARGET_AVX512CD.
4581 2014-02-26 Richard Biener <rguenther@suse.de>
4584 * ipa.c (walk_polymorphic_call_targets): Properly guard
4585 call to inline_update_overall_summary.
4587 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4590 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4591 and latches only if requested. Fix latch if it is removed.
4592 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4593 LOOPS_HAVE_PREHEADERS.
4595 2014-02-25 Andrew Pinski <apinski@cavium.com>
4597 * builtins.c (expand_builtin_thread_pointer): Create a new target
4598 when the target is NULL.
4600 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4602 PR rtl-optimization/60317
4603 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4604 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4605 * lra-assigns.c: Include params.h.
4606 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4607 other reload pseudos considerations.
4609 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4611 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4612 to use canonical form for nor<mode>3.
4614 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4617 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4620 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4622 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4623 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4624 (ix86_handle_option): Handle OPT_mprefetchwt1.
4625 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4626 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4628 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4629 OPTION_MASK_ISA_PREFETCHWT1.
4630 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4631 (PTA_PREFETCHWT1): New.
4632 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4633 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4634 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4635 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4636 (*prefetch_avx512pf_<mode>_: Change into ...
4637 (*prefetch_prefetchwt1_<mode>: This.
4638 * config/i386/i386.opt (mprefetchwt1): New.
4639 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4640 (_mm_prefetch): Handle intent to write.
4641 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4643 2014-02-25 Richard Biener <rguenther@suse.de>
4646 * emit-rtl.c (mem_attrs_htab): Remove.
4647 (mem_attrs_htab_hash): Likewise.
4648 (mem_attrs_htab_eq): Likewise.
4649 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4650 (init_emit_once): Do not allocate mem_attrs_htab.
4652 2014-02-25 Richard Biener <rguenther@suse.de>
4655 * lto-opts.c (lto_write_options): Output non-explicit conservative
4656 -fwrapv, -fno-trapv and -fno-strict-overflow.
4657 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4658 (run_gcc): And pass them through.
4660 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4662 * sel-sched.c (calculate_new_fences): New parameter ptime.
4663 Calculate it as a maximum over all fence cycles.
4664 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4665 Print the final schedule timing when sched_verbose.
4667 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4669 PR rtl-optimization/60292
4670 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4671 bit fot the fence instruction.
4673 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4675 * calls.h: Fix typo in comment.
4677 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4679 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4680 adjusting offsetable addresses.
4682 2014-02-24 Guozhi Wei <carrot@google.com>
4684 * sparseset.h (sparseset_pop): Fix the wrong index.
4686 2014-02-24 Walter Lee <walt@tilera.com>
4688 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4689 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4691 * common/config/tilegx/tilegx-common.c
4692 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4693 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4695 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4696 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4697 (tilegx_gimplify_va_arg_expr): Handle big endian.
4698 (tilegx_expand_unaligned_load): Ditto.
4699 (tilegx_expand_unaligned_store): Ditto.
4700 (TARGET_RETURN_IN_MSB): New.
4701 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4702 (TARGET_ENDIAN_DEFAULT): New.
4703 (TARGET_BIG_ENDIAN): Handle big endian.
4704 (BYTES_BIG_ENDIAN): Ditto.
4705 (WORDS_BIG_ENDIAN): Ditto.
4706 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4709 * config/tilegx/tilegx.md (extv): Handle big endian.
4711 (insn_st<n>): Ditto.
4712 (insn_st<n>_add<bitsuffix>): Ditto.
4713 (insn_stnt<n>): Ditto.
4714 (insn_stnt<n>_add<bitsuffix>):Ditto.
4715 (vec_interleave_highv8qi): Handle big endian.
4716 (vec_interleave_highv8qi_be): New.
4717 (vec_interleave_highv8qi_le): New.
4718 (insn_v1int_h): Handle big endian.
4719 (vec_interleave_lowv8qi): Handle big endian.
4720 (vec_interleave_lowv8qi_be): New.
4721 (vec_interleave_lowv8qi_le): New.
4722 (insn_v1int_l): Handle big endian.
4723 (vec_interleave_highv4hi): Handle big endian.
4724 (vec_interleave_highv4hi_be): New.
4725 (vec_interleave_highv4hi_le): New.
4726 (insn_v2int_h): Handle big endian.
4727 (vec_interleave_lowv4hi): Handle big endian.
4728 (vec_interleave_lowv4hi_be): New.
4729 (vec_interleave_lowv4hi_le): New.
4730 (insn_v2int_l): Handle big endian.
4731 (vec_interleave_highv2si): Handle big endian.
4732 (vec_interleave_highv2si_be): New.
4733 (vec_interleave_highv2si_le): New.
4734 (insn_v4int_h): Handle big endian.
4735 (vec_interleave_lowv2si): Handle big endian.
4736 (vec_interleave_lowv2si_be): New.
4737 (vec_interleave_lowv2si_le): New.
4738 (insn_v4int_l): Handle big endian.
4739 * config/tilegx/tilegx.opt (mbig-endian): New option.
4740 (mlittle-endian): New option.
4741 * doc/install.texi: Document tilegxbe-linux.
4742 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4744 2014-02-24 Martin Jambor <mjambor@suse.cz>
4747 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4748 there are no parameter descriptors.
4750 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4752 PR rtl-optimization/60268
4753 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4754 initialization to ...
4755 (sched_rgn_init): ... here.
4756 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4758 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4760 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4763 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4765 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4768 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4770 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4771 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4773 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4775 * config/microblaze/predicates.md: Add cmp_op predicate.
4776 * config/microblaze/microblaze.md: Add branch_compare instruction
4777 which uses cmp_op predicate and emits cmp insn before branch.
4778 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4779 to microblaze_expand_conditional_branch and consolidate logic.
4780 (microblaze_expand_conditional_branch): emit branch_compare
4781 insn instead of handling cmp op separate from branch insn.
4783 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4785 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4788 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4790 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4791 define_insn with define_expand and new define_insn
4792 *altivec_lve<VI_char>x_internal.
4793 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4794 and new define_insn *altivec_stve<VI_char>x_internal.
4795 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4797 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4799 (altivec_expand_stvex_be): New function.
4801 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4803 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4804 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4805 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4806 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4808 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4811 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4812 instead of emit_move_insn.
4814 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4816 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4818 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4819 gen_altivec_vsumsws.
4821 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4823 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4824 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4825 (altivec_lvxl_<mode>): New define_expand incorporating
4826 -maltivec=be semantics where needed.
4827 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4828 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4829 semantics where needed.
4830 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4831 (altivec_stvx_<mode>): New define_expand incorporating
4832 -maltivec=be semantics where needed.
4833 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4834 VM2 iterator instead of V4SI.
4835 (altivec_stvxl_<mode>): New define_expand incorporating
4836 -maltivec=be semantics where needed.
4837 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4838 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4839 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4840 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4841 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4842 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4843 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4844 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4845 ALTIVEC_BUILTIN_STVXL.
4846 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4847 (altivec_expand_stvx_be): Likewise.
4848 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4849 (altivec_expand_lvx_be): Likewise.
4850 (altivec_expand_stvx_be): Likewise.
4851 (altivec_expand_builtin): Add cases for
4852 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4853 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4854 (altivec_init_builtins): Add definitions for
4855 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4856 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4858 2014-02-21 Catherine Moore <clm@codesourcery.com>
4860 * doc/invoke.texi (mvirt, mno-virt): Document.
4861 * config/mips/mips.opt (mvirt): New option.
4862 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4864 2014-02-21 Richard Biener <rguenther@suse.de>
4866 PR tree-optimization/60276
4867 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4868 (STMT_VINFO_MIN_NEG_DIST): New macro.
4869 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4870 STMT_VINFO_MIN_NEG_DIST.
4871 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4872 made for negative dependence distances still hold.
4874 2014-02-21 Richard Biener <rguenther@suse.de>
4877 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4878 DECL_INITIAL for globals not in the current function context.
4880 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4882 PR tree-optimization/56490
4883 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4884 * tree-ssa-uninit.c: Include params.h.
4885 (compute_control_dep_chain): Add num_calls argument, return false
4886 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4887 num_calls to recursive call.
4888 (find_predicates): Change dep_chain into normal array,
4889 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4890 variable and adjust compute_control_dep_chain caller.
4891 (find_def_preds): Likewise.
4893 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4895 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4896 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4898 2014-02-21 Nick Clifton <nickc@redhat.com>
4900 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4901 (pushhi1): Likewise.
4902 (popqi1): Add mode to pre_dec.
4905 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4907 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4908 mode for mask of V8SFmode permutation.
4910 2014-02-20 Richard Henderson <rth@redhat.com>
4913 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4914 a new pseudo for OLDVAL.
4916 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4919 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4920 gen_reg_rtx if d->testing_p.
4921 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4922 if d->testing_p and we will certainly return true.
4923 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4926 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4928 * emit-rtl.c (gen_reg_rtx): Assert that
4929 crtl->emit.regno_pointer_align_length is non-zero.
4931 2014-02-20 Richard Henderson <rth@redhat.com>
4934 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4935 on failure the store back into EXPECT.
4937 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4938 Sandra Loosemore <sandra@codesourcery.com>
4940 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
4941 * config/nios2/nios2.c (nios2_function_profiler): Add
4942 -fPIC (flag_pic == 2) support.
4943 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4944 (nios2_large_offset_p): New function.
4945 (nios2_unspec_reloc_p): Move up position, update to use
4946 nios2_large_offset_p.
4947 (nios2_unspec_address): Remove function.
4948 (nios2_unspec_offset): New function.
4949 (nios2_large_got_address): New function.
4950 (nios2_got_address): Add large offset support.
4951 (nios2_legitimize_tls_address): Update usage of removed and new
4953 (nios2_symbol_binds_local_p): New function.
4954 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
4955 (nios2_legitimize_address): Update to use nios2_large_offset_p.
4956 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
4957 (nios2_print_operand): Merge H/L processing, add hiadj/lo
4958 processing for (const (unspec ...)).
4959 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
4961 2014-02-20 Richard Biener <rguenther@suse.de>
4963 * tree-cfg.c (replace_uses_by): Mark altered BBs before
4964 doing the substitution.
4965 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
4967 2014-02-20 Martin Jambor <mjambor@suse.cz>
4970 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
4971 info when checking whether lattices are bottom.
4973 2014-02-20 Richard Biener <rguenther@suse.de>
4976 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
4979 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
4982 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
4983 parameter specifying the scaling.
4984 (inline_call): Update.
4985 (want_inline_recursively): Guard division by zero.
4986 (recursive_inlining): Update.
4987 * ipa-inline.h (clone_inlined_nodes): Update.
4989 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4992 * config/i386/i386.c (classify_argument): Pass structures of size
4993 64 bytes or less in register.
4995 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4996 Kirill Yukhin <kirill.yukhin@intel.com>
4998 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
4999 (_mm_rcp28_round_ss): Ditto.
5000 (_mm_rsqrt28_round_sd): Ditto.
5001 (_mm_rsqrt28_round_ss): Ditto.
5002 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
5003 (_mm_rcp14_round_ss): Ditto.
5004 (_mm_rsqrt14_round_sd): Ditto.
5005 (_mm_rsqrt14_round_ss): Ditto.
5006 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
5007 the first input operand, get rid of match_dup.
5008 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
5010 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
5012 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
5013 operand as the first input operand, set type attribute.
5014 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
5016 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
5017 operand as the first input operand, set type attribute.
5019 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5021 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
5024 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
5027 * config/i386/i386.c (construct_container): Remove TFmode check
5028 for X86_64_INTEGER_CLASS.
5030 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
5033 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
5034 only when -Wpsabi is enabled.
5036 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
5039 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
5040 passing arrays in registers are the same as for structs, so remove the
5041 special case for them.
5043 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
5045 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
5046 destination type, extract only the valid bits if the source type is not
5047 integral and has a different mode.
5049 2014-02-19 Richard Biener <rguenther@suse.de>
5052 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
5055 2014-02-19 Richard Biener <rguenther@suse.de>
5058 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
5059 (ipa_modify_call_arguments): Emit an argument load explicitely and
5060 preserve virtual SSA form there and for the replacement call.
5061 Do not update SSA form nor free dominance info.
5063 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5065 * ipa.c (function_and_variable_visibility): Also clear WEAK
5066 flag when disolving COMDAT_GROUP.
5068 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5070 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
5071 * ipa-prop.c (ipa_set_jf_known_type): Return early when
5073 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
5074 do more sanity checks.
5075 (detect_type_change): Return true when giving up early.
5076 (compute_complex_assign_jump_func): Fix type parameter of
5077 ipa_set_ancestor_jf.
5078 (compute_complex_ancestor_jump_func): Likewise.
5079 (update_jump_functions_after_inlining): Fix updating of
5081 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
5083 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5085 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
5086 inline clones when edge disappears.
5088 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
5091 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
5092 Split 64-bit moves into 2 patterns. Do not allow the use of
5093 direct move for TDmode in little endian, since the decimal value
5094 has little endian bytes within a word, but the 64-bit pieces are
5095 ordered in a big endian fashion, and normal subreg's of TDmode are
5097 (mov<mode>_64bit_dm): Likewise.
5098 (movtd_64bit_nodm): Likewise.
5100 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5102 PR tree-optimization/60174
5103 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
5104 statement of an SSA_NAME that occurs in an abnormal PHI node.
5106 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5109 * final.c (SEEN_BB): Remove.
5110 (SEEN_NOTE, SEEN_EMITTED): Renumber.
5111 (final_scan_insn): Don't force_source_line on second
5112 NOTE_INSN_BASIC_BLOCK.
5114 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
5117 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
5118 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
5119 (type_natural_mode): Warn ABI change when %zmm register is not
5120 available for AVX512F vector value passing.
5122 2014-02-18 Kai Tietz <ktietz@redhat.com>
5125 * config/i386/i386.c (ix86_expand_prologue): Use value in
5126 rax register as displacement when restoring %r10 or %rax.
5127 Fix wrong offset when restoring both registers.
5129 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5131 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
5132 assertion with conditional return.
5134 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5135 Uros Bizjak <ubizjak@gmail.com>
5138 * config/i386/driver-i386.c (host_detect_local_cpu): If
5139 YMM state is not saved by the OS, also clear has_f16c. Move
5140 CPUID 0x80000001 handling before YMM state saving checking.
5142 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
5144 PR rtl-optimization/58960
5145 * haifa-sched.c (alloc_global_sched_pressure_data): New,
5146 factored out from ...
5147 (sched_init): ... here.
5148 (free_global_sched_pressure_data): New, factored out from ...
5149 (sched_finish): ... here.
5150 * sched-int.h (free_global_sched_pressure_data): Declare.
5151 * sched-rgn.c (nr_regions_initial): New static global.
5152 (haifa_find_rgns): Initialize it.
5153 (schedule_region): Disable sched-pressure for the newly
5156 2014-02-17 Richard Biener <rguenther@suse.de>
5158 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
5159 release SSA defs of pattern stmts.
5161 2014-02-17 Richard Biener <rguenther@suse.de>
5163 * tree-inline.c (expand_call_inline): Release the virtual
5164 operand defined by the call we are about to inline.
5166 2014-02-17 Richard Biener <rguenther@suse.de>
5168 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
5170 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
5171 Ilya Tocar <ilya.tocar@intel.com>
5173 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
5174 arguments order in builtin.
5175 (_mm512_permutexvar_epi64): Ditto.
5176 (_mm512_mask_permutexvar_epi64): Ditto
5177 (_mm512_maskz_permutexvar_epi32): Ditto
5178 (_mm512_permutexvar_epi32): Ditto
5179 (_mm512_mask_permutexvar_epi32): Ditto
5181 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5183 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
5184 (p8_vmrgow): Likewise.
5186 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5188 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
5191 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
5194 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
5195 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
5196 into 64-bit and 32-bit moves. On 64-bit moves, add support for
5197 using direct move instructions on ISA 2.07. Also adjust
5198 instruction length for 64-bit.
5199 (mov<mode>_64bit, TFmode/TDmode): Likewise.
5200 (mov<mode>_32bit, TFmode/TDmode): Likewise.
5202 2014-02-15 Alan Modra <amodra@gmail.com>
5206 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
5207 find_replacement on parts of insn rtl that might be reloaded.
5209 2014-02-15 Richard Biener <rguenther@suse.de>
5211 PR tree-optimization/60183
5212 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
5213 (tree_ssa_phiprop): Calculate and free post-dominators.
5215 2014-02-14 Jeff Law <law@redhat.com>
5217 PR rtl-optimization/60131
5218 * ree.c (get_extended_src_reg): New function.
5219 (combine_reaching_defs): Use it rather than assuming location of REG.
5220 (find_and_remove_re): Verify first operand of extension is
5221 a REG before adding the insns to the copy list.
5223 2014-02-14 Roland McGrath <mcgrathr@google.com>
5225 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
5226 * configure: Regenerated.
5227 * config.in: Regenerated.
5228 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
5229 instead of ASM_SHORT.
5231 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
5232 Richard Earnshaw <rearnsha@arm.com>
5234 PR rtl-optimization/59535
5235 * lra-constraints.c (process_alt_operands): Encourage alternative
5236 when unassigned pseudo class is superset of the alternative class.
5237 (inherit_reload_reg): Don't inherit when optimizing for code size.
5238 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
5239 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
5240 modes not less than 4 for Thumb1.
5242 2014-02-14 Kyle McMartin <kyle@redhat.com>
5245 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
5247 2014-02-14 Richard Biener <rguenther@suse.de>
5249 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
5250 (get_frame_arg): Drop the assert with langhook types_compatible_p.
5251 Do not strip INDIRECT_REFs.
5253 2014-02-14 Richard Biener <rguenther@suse.de>
5256 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5257 DECL_FUNCTION_SPECIFIC_TARGET.
5258 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
5259 * tree-streamer-out.c (pack_ts_target_option): Remove.
5260 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
5261 (write_ts_function_decl_tree_pointers): Do not stream
5262 DECL_FUNCTION_SPECIFIC_TARGET.
5263 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5264 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5265 (lto_input_ts_function_decl_tree_pointers): Do not stream
5266 DECL_FUNCTION_SPECIFIC_TARGET.
5268 2014-02-14 Jakub Jelinek <jakub@redhat.com>
5270 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5271 (get_initial_def_for_induction, vectorizable_induction): Ignore
5272 debug stmts when looking for exit_phi.
5273 (vectorizable_live_operation): Fix up condition.
5275 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5277 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5278 nreverse() because it changes the content of original tree list.
5280 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5282 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5283 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5285 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5287 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5288 GNU coding standards.
5290 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5293 * dwarf2out.c (gen_subprogram_die): Don't call
5294 add_calling_convention_attribute if subr_die is old_die.
5296 2014-02-13 Sharad Singhai <singhai@google.com>
5298 * doc/optinfo.texi: Fix order of nodes.
5300 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5302 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5303 operands[2], not operands[3].
5305 2014-02-13 Richard Biener <rguenther@suse.de>
5308 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5309 mention the possibility of an in-tree build.
5310 (CLooG): Update recommended version to 0.18.1, mention the
5311 possibility of an in-tree build and clarify that the ISL
5312 bundled with CLooG does not work.
5314 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5317 * expr.c (compress_float_constant): If x is a hard register,
5318 extend into a pseudo and then move to x.
5320 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5322 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5323 caused by bad second argument to warning_at() with -mhotpatch and
5324 nested functions (e.g. with gfortran).
5326 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5328 * opts.c (option_name): Remove "enabled by default" rider.
5330 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5332 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5334 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5335 Uros Bizjak <ubizjak@gmail.com>
5338 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5339 * configure: Regenerated.
5341 2014-02-12 Richard Biener <rguenther@suse.de>
5343 * vec.c (vec_prefix::calculate_allocation): Move as
5344 inline variant to vec.h.
5345 (vec_prefix::calculate_allocation_1): New out-of-line version.
5346 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5347 (vec_prefix::m_has_auto_buf): Rename to ...
5348 (vec_prefix::m_using_auto_storage): ... this.
5349 (vec_prefix::calculate_allocation): Inline the easy cases
5350 and dispatch to calculate_allocation_1 which doesn't need the
5352 (va_heap::reserve): Use gcc_checking_assert.
5353 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5354 m_using_auto_storage.
5355 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5357 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5358 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5359 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5361 2014-02-12 Richard Biener <rguenther@suse.de>
5363 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5364 when we found a dependence.
5366 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5368 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5370 (maybe_fold_stmt): ... into this new function.
5371 * omp-low.c (lower_omp): Update comment.
5373 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5376 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5379 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5381 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5382 identifiers in comments.
5383 (cortexa53_extra_costs): Likewise.
5384 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5385 (cortexa7_extra_costs): Likewise.
5386 (cortexa12_extra_costs): Likewise.
5387 (cortexa15_extra_costs): Likewise.
5388 (v7m_extra_costs): Likewise.
5390 2014-02-12 Richard Biener <rguenther@suse.de>
5393 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5394 of posix_memalign being successful.
5395 (lower_stmt): Restrict lowering of posix_memalign to when
5396 -ftree-bit-ccp is enabled.
5398 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5400 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5402 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5404 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5406 PR rtl-optimization/60116
5407 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5408 other_insn once the combination has been validated.
5410 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5413 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5415 * ipa-devirt.c: Include demangle.h
5416 (odr_violation_reported): New static variable.
5417 (add_type_duplicate): Update odr_violations.
5418 (maybe_record_node): Add completep parameter; update it.
5419 (record_target_from_binfo): Add COMPLETEP parameter;
5420 update it as needed.
5421 (possible_polymorphic_call_targets_1): Likewise.
5422 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5423 rename FINAL to COMPLETE.
5424 (record_targets_from_bases): Sanity check we found the binfo;
5425 fix COMPLETEP updating.
5426 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5427 parameter, fix computing of COMPLETEP.
5428 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5429 at LTO time do demangling.
5430 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5431 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5433 (gimple_get_virt_method_for_binfo): Likewise.
5434 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5435 gimple_get_virt_method_for_vtable): Update prototypes.
5437 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5440 * genautomata.c (add_presence_absence): Fix typo with
5441 {final_}presence_list.
5443 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5446 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5447 for VSX/Altivec vectors that land in GPR registers.
5449 2014-02-11 Richard Henderson <rth@redhat.com>
5450 Jakub Jelinek <jakub@redhat.com>
5453 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5454 around drhs if type conversion to lacc->type is not useless.
5456 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5458 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5460 (cortex-a57.cortex-a53): Likewise.
5461 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5463 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5465 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5468 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5470 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5471 add_options_for_arm_vfp3.
5473 2014-02-11 Jeff Law <law@redhat.com>
5476 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5477 object with an undesirable mode.
5479 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5482 * config/i386/sol2-9.h: New file.
5483 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5484 *-*-solaris2.9*): Use it.
5486 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5488 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5489 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5491 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5493 * config/microblaze/microblaze.c: Extend mcpu version format
5495 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5497 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5499 2014-02-10 Richard Henderson <rth@redhat.com>
5502 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5503 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5504 ms-abi vs -mno-accumulate-outgoing-args.
5505 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5506 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5509 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5512 * cfgexpand.c (expand_asm_operands): Attach source location to
5513 ASM_INPUT rtx objects.
5514 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5516 2014-02-10 Nick Clifton <nickc@redhat.com>
5518 * config/mn10300/mn10300.c (popcount): New function.
5519 (mn10300_expand_prologue): Include saved registers in stack usage
5522 2014-02-10 Jeff Law <law@redhat.com>
5525 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5526 when changing the SET_DEST of a prior insn to avoid an input reload.
5528 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5530 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5531 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5532 -mcall-openbsd, or -mcall-linux.
5533 (CC1_ENDIAN_BIG_SPEC): Remove.
5534 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5535 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5536 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5537 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5538 and %cc1_endian_default.
5539 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5541 2014-02-10 Richard Biener <rguenther@suse.de>
5543 PR tree-optimization/60115
5544 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5545 MEM_REF handling. Properly verify that the accesses are not
5546 out of the objects bound.
5548 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5550 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5553 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5555 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5556 proper constants and fix formatting.
5557 (possible_polymorphic_call_targets): Fix formatting.
5559 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5560 Ilya Tocar <ilya.tocar@intel.com>
5562 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5563 (_mm512_loadu_epi32): Renamed into...
5564 (_mm512_loadu_si512): This.
5565 (_mm512_storeu_epi32): Renamed into...
5566 (_mm512_storeu_si512): This.
5567 (_mm512_maskz_ceil_ps): Removed.
5568 (_mm512_maskz_ceil_pd): Ditto.
5569 (_mm512_maskz_floor_ps): Ditto.
5570 (_mm512_maskz_floor_pd): Ditto.
5571 (_mm512_floor_round_ps): Ditto.
5572 (_mm512_floor_round_pd): Ditto.
5573 (_mm512_ceil_round_ps): Ditto.
5574 (_mm512_ceil_round_pd): Ditto.
5575 (_mm512_mask_floor_round_ps): Ditto.
5576 (_mm512_mask_floor_round_pd): Ditto.
5577 (_mm512_mask_ceil_round_ps): Ditto.
5578 (_mm512_mask_ceil_round_pd): Ditto.
5579 (_mm512_maskz_floor_round_ps): Ditto.
5580 (_mm512_maskz_floor_round_pd): Ditto.
5581 (_mm512_maskz_ceil_round_ps): Ditto.
5582 (_mm512_maskz_ceil_round_pd): Ditto.
5583 (_mm512_expand_pd): Ditto.
5584 (_mm512_expand_ps): Ditto.
5585 * config/i386/i386.c (ix86_builtins): Remove
5586 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5587 (bdesc_args): Ditto.
5588 * config/i386/predicates.md (const1256_operand): New.
5589 (const_1_to_2_operand): Ditto.
5590 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5591 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5592 (*avx512pf_gatherpf<mode>sf): Ditto.
5593 (avx512pf_gatherpf<mode>df): Ditto.
5594 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5595 (*avx512pf_gatherpf<mode>df): Ditto.
5596 (avx512pf_scatterpf<mode>sf): Ditto.
5597 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5598 (*avx512pf_scatterpf<mode>sf): Ditto.
5599 (avx512pf_scatterpf<mode>df): Ditto.
5600 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5601 (*avx512pf_scatterpf<mode>df): Ditto.
5602 (avx512f_expand<mode>): Removed.
5603 (<shift_insn><mode>3<mask_name>): Change predicate type.
5605 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5607 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5608 not at the end of datarefs vector use ordered_remove to avoid
5609 reordering datarefs vector.
5612 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5613 mark local addressable non-static vars as GOVD_PRIVATE
5614 instead of GOVD_LOCAL.
5615 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5616 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5620 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5621 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5622 assume_aligned or alloc_align attributes.
5623 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5624 arguments. Handle also assume_aligned and alloc_align attributes.
5625 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5626 calls to functions with assume_aligned or alloc_align attributes.
5627 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5629 2014-02-08 Terry Guo <terry.guo@arm.com>
5631 * doc/invoke.texi: Document ARM -march=armv7e-m.
5633 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5635 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5636 flag on __cilkrts_rethrow builtin.
5639 * ipa-cp.c (determine_versionability): Fail at -O0
5640 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5641 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5644 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5647 * tree-inline.c (copy_forbidden): Fail for
5648 __attribute__((optimize (0))) functions.
5650 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5652 * varpool.c: Include pointer-set.h.
5653 (varpool_remove_unreferenced_decls): Variables in other partitions
5654 will not be output; be however careful to not lose information
5657 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5659 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5660 lookup in the vtable constructor.
5662 2014-02-07 Jeff Law <law@redhat.com>
5665 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5666 define_insn_and_split.
5668 * ipa-inline.c (inline_small_functions): Fix typos.
5670 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5672 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5673 (s390_can_use_return_insn): Declare.
5674 * config/s390/s390.h (EPILOGUE_USES): Define.
5675 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5677 (s390_chunkify_start): Handle return JUMP_LABELs.
5678 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5679 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5680 (s390_can_use_return_insn): New functions.
5681 (s390_fix_long_loop_prediction): Handle conditional returns.
5682 (TARGET_SET_UP_BY_PROLOGUE): Define.
5683 * config/s390/s390.md (ANY_RETURN): New code iterator.
5684 (*creturn, *csimple_return, return, simple_return): New patterns.
5686 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5688 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5689 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5690 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5691 REG_CFA_RESTORE list when deciding not to restore a register.
5693 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5695 * config/s390/s390.c: Include tree-pass.h and context.h.
5696 (s390_early_mach): New function, split out from...
5697 (s390_emit_prologue): ...here.
5698 (pass_data_s390_early_mach): New pass structure.
5699 (pass_s390_early_mach): New class.
5700 (s390_option_override): Create and register early_mach pass.
5701 Move to end of file.
5703 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5705 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5706 to match for the exit block.
5708 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5710 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5711 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5712 Reject misaligned operands.
5714 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5716 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5718 2014-02-07 Richard Biener <rguenther@suse.de>
5721 * gimple-low.c (lower_builtin_posix_memalign): New function.
5722 (lower_stmt): Call it to lower posix_memalign in a way
5723 to make alignment info accessible.
5725 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5728 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5729 __builtin_setjmp_receiver.
5731 2014-02-07 Richard Biener <rguenther@suse.de>
5734 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5735 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5736 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5737 Handle BUILT_IN_POSIX_MEMALIGN.
5738 (find_func_clobbers): Likewise.
5739 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5740 (call_may_clobber_ref_p_1): Likewise.
5742 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5745 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5748 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5751 * lto-cgraph.c (lto_output_node): Use
5752 symtab_get_symbol_partitioning_class.
5753 (lto_output_varpool_node): likewise.
5754 (symtab_get_symbol_partitioning_class): Move here from
5756 * cgraph.h (symbol_partitioning_class): Likewise.
5757 (symtab_get_symbol_partitioning_class): Declare.
5759 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5761 * ggc.h (ggc_internal_cleared_alloc): New macro.
5762 * vec.h (vec_safe_copy): Handle memory stats.
5763 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5764 * target-globals.c (save_target_globals): Likewise.
5766 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5769 * expr.c (emit_move_resolve_push): Export; be bit more selective
5770 on when to clear alias set.
5771 * expr.h (emit_move_resolve_push): Declare.
5772 * function.h (struct function): Add tail_call_marked.
5773 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5774 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5775 * config/i386/i386.md (TImode move expander): De not call
5777 (FP push expanders): Preserve memory attributes.
5778 * config/i386/sse.md (push<mode>1): Remove.
5779 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5780 (ix86_expand_push): Remove.
5781 * config/i386/mmx.md (push<mode>1): Remove.
5783 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5785 PR rtl-optimization/60030
5786 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5787 lopart with paradoxical subreg before shifting it up by hprec.
5789 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5791 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5792 Remove extra newline at end of file.
5793 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5794 (arm_issue_rate): Handle cortexa57.
5795 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5796 (cortex-a57.cortex-a53): Likewise.
5798 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5801 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5802 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5804 (arm_expand_prologue): Adjust all callers.
5805 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5806 info, registers also at the lowest numbered registers side. Use
5807 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5811 * var-tracking.c (adjust_mems): Before adding a SET to
5812 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5814 2014-02-06 Alan Modra <amodra@gmail.com>
5817 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5818 change SDmode to DDmode when lra_in_progress.
5820 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5823 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5824 free_data_ref on the dr first, and before goto again also set dr
5825 to the next dr. For simd_lane_access, free old datarefs[i] before
5826 overwriting it. For get_vectype_for_scalar_type failure, don't
5827 free_data_ref if simd_lane_access.
5829 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5832 * tree.h (opts_for_fn): New inline function.
5833 (opt_for_fn): Define.
5834 * config/i386/i386.c (ix86_function_regparm): Use
5835 opt_for_fn (decl, optimize) instead of optimize.
5837 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5839 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5840 for SYMBOL_REF in large memory model.
5842 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5844 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5846 (cortex-a57): Likewise.
5847 (cortex-a57.cortex-a53): Likewise.
5849 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5850 Kugan Vivekanandarajah <kuganv@linaro.org>
5852 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5854 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5856 2014-02-06 Richard Biener <rguenther@suse.de>
5858 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5859 set_loop_copy and initialize_original_copy_tables.
5861 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5863 * config/aarch64/aarch64-simd.md
5864 (aarch64_ashr_simddi): Change QI to SI.
5866 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5867 Jakub Jelinek <jakub@redhat.com>
5870 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5873 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5875 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5876 CODE_FOR_altivec_vpku[hw]um to
5877 CODE_FOR_altivec_vpku[hw]um_direct.
5878 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5879 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5880 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5881 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5883 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5885 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5886 generation for -maltivec=be.
5887 (altivec_vsumsws): Simplify redundant test.
5889 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5891 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5892 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5893 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5894 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5895 gen_altivec_vpkuwum.
5896 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5898 (altivec_vpks<VI_char>ss): Likewise.
5899 (altivec_vpks<VI_char>us): Likewise.
5900 (altivec_vpku<VI_char>us): Likewise.
5901 (altivec_vpku<VI_char>um): Likewise.
5902 (altivec_vpku<VI_char>um_direct): New (copy of
5903 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5905 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5906 target is little endian and -maltivec=be is not specified.
5907 (*altivec_vupkhs<VU_char>_direct): New (copy of
5908 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5909 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5910 target is little endian and -maltivec=be is not specified.
5911 (*altivec_vupkls<VU_char>_direct): New (copy of
5912 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5913 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5914 little endian and -maltivec=be is not specified.
5915 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5916 little endian and -maltivec=be is not specified.
5918 2014-02-05 Richard Henderson <rth@redhat.com>
5921 * combine-stack-adj.c: Revert r206943.
5922 * sched-int.h (struct deps_desc): Add last_args_size.
5923 * sched-deps.c (init_deps): Initialize it.
5924 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5925 contain REG_ARGS_SIZE notes.
5927 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5929 * lto-cgraph.c (asm_nodes_output): Make global.
5930 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5931 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5932 (driver_handle_option): Handle OPT_fwpa.
5934 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5937 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5938 a comment typo and formatting issue. If odr_hash hasn't been
5939 created, return vNULL and set *completep to false.
5942 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5943 bb with no successors.
5945 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5948 * doc/invoke.texi (-march): Clarify documentation for ARM.
5952 2014-02-05 Richard Biener <rguenther@suse.de>
5954 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
5955 when not vectorizing because of too many alias checks.
5956 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
5957 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
5959 2014-02-05 Nick Clifton <nickc@redhat.com>
5961 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
5962 accept extended registers in any mode when compiling for the MN10300.
5964 2014-02-05 Yury Gribov <y.gribov@samsung.com>
5966 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
5967 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
5968 sanitization attributes.
5969 (can_inline_edge_p): Likewise.
5970 (sanitize_attrs_match_for_inline_p): New function.
5972 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5974 * ipa-prop.c (detect_type_change): Shor circuit testing of
5975 type changes on THIS pointer.
5977 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
5980 * config/pa/pa.c (legitimize_tls_address): Return original address
5981 if not passed a SYMBOL_REF rtx.
5982 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
5984 (pa_emit_move_sequence): Simplify TLS source operands.
5985 (pa_legitimate_constant_p): Reject all TLS constants.
5986 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
5987 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
5989 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5991 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
5992 groups when we know they are controlled by LTO.
5993 * varasm.c (default_binds_local_p_1): If object is in other partition,
5994 it will be resolved locally.
5996 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5998 * config/host-linux.c (linux_gt_pch_use_address): Don't
5999 use SSIZE_MAX because it is not always defined.
6001 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
6004 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
6005 threshold for pseudo splitting.
6006 (update_ebb_live_info): Process call argument hard registers and
6007 hard registers from insn definition too.
6008 (max_small_class_regs_num): New constant.
6009 (inherit_in_ebb): Update live hard regs through EBBs. Update
6010 reloads_num only for small register classes. Don't split for
6013 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
6016 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
6019 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6021 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
6024 2014-02-04 Marek Polacek <polacek@redhat.com>
6026 * gdbinit.in (pel): Define.
6028 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6030 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
6033 2014-02-04 Richard Biener <rguenther@suse.de>
6036 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
6037 in function context local.
6038 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
6039 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
6040 similar to LTO_imported_decl_ref.
6042 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6044 PR tree-optimization/60002
6045 * cgraphclones.c (build_function_decl_skip_args): Clear
6048 PR tree-optimization/60023
6049 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
6050 false to gsi_replace.
6051 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
6052 has been in some EH region and vec_stmt could throw, add
6053 vec_stmt into the same EH region.
6054 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
6055 has no lhs, ignore it.
6056 * internal-fn.c (expand_MASK_LOAD): Likewise.
6059 * tree-inline.c (copy_forbidden): Fail for
6060 __attribute__((optimize (0))) functions.
6063 * omp-low.c (simd_clone_struct_copy): If from->inbranch
6064 is set, copy one less argument.
6065 (expand_simd_clones): Don't subtract clone_info->inbranch
6066 from simd_clone_struct_alloc argument.
6068 PR rtl-optimization/57915
6069 * recog.c (simplify_while_replacing): If all unary/binary/relational
6070 operation arguments are constant, attempt to simplify those.
6073 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
6074 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
6076 2014-02-04 Richard Biener <rguenther@suse.de>
6078 PR tree-optimization/60012
6079 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
6080 TBAA disambiguation to all DDRs.
6082 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6085 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
6086 (LINK_SPEC): Use it for -shared, -shared-libgcc.
6088 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6091 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
6093 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6095 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
6096 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
6098 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6101 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
6102 to figure out targets of polymorphic calls with known decl.
6103 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6104 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
6105 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
6106 (get_polymorphic_call_info): ... here.
6107 (get_polymorphic_call_info_from_invariant): New function.
6109 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6111 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
6112 lookup via vtable pointer; check for type consistency
6113 and turn inconsitent facts into UNREACHABLE.
6114 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6115 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
6116 type inconsistent querries; return UNREACHABLE instead.
6118 2014-02-03 Richard Henderson <rth@twiddle.net>
6121 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
6122 already processed this node.
6123 (normalize_one_pred_1): Pass along mark_set.
6124 (normalize_one_pred): Create and destroy a pointer_set_t.
6125 (normalize_one_pred_chain): Likewise.
6127 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
6129 PR gcov-profile/58602
6130 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
6132 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6135 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
6136 -fno-devirtualize; try to devirtualize by the knowledge of
6137 virtual table pointer given by aggregate propagation.
6138 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6139 (ipa_print_node_jump_functions): Dump also offset that
6140 is relevant for polymorphic calls.
6141 (determine_known_aggregate_parts): Add arg_type parameter; use it
6142 instead of determining the type from pointer type.
6143 (ipa_compute_jump_functions_for_edge): Update call of
6144 determine_known_aggregate_parts.
6145 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
6146 (gimple_get_virt_method_for_binfo): ... here; simplify using
6147 vtable_pointer_value_to_vtable.
6148 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
6149 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
6150 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
6151 (vtable_pointer_value_to_vtable): Break out from ...; handle also
6153 (vtable_pointer_value_to_binfo): ... here.
6154 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
6156 2014-02-03 Teresa Johnson <tejohnson@google.com>
6158 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
6159 redef of outer loop index variable.
6161 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
6165 * doc/extend.texi (Function Attributes): Typo.
6167 2014-02-03 Cong Hou <congh@google.com>
6169 PR tree-optimization/60000
6170 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
6171 if the vectorized statement is a store. A store statement can only
6172 appear at the end of pattern statements.
6174 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6176 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
6177 (ix86_option_override_internal): Default long double to 64-bit for
6178 32-bit Bionic and to 128-bit for 64-bit Bionic.
6180 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
6181 TARGET_LONG_DOUBLE_128 is true.
6182 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
6184 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
6185 (mlong-double-64): Negate -mlong-double-128.
6186 (mlong-double-128): New option.
6188 * config/i386/i386-c.c (ix86_target_macros): Define
6189 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
6191 * doc/invoke.texi: Document -mlong-double-128.
6193 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6195 PR rtl-optimization/60024
6196 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
6198 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
6200 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
6202 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6204 PR rtl-optimization/57662
6205 * sel-sched.c (code_motion_path_driver): Do not mark already not
6206 existing blocks in the visiting bitmap.
6208 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6210 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
6211 on the insn being emitted.
6213 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
6214 Will Deacon <will.deacon@arm.com>
6216 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
6218 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6220 * config/arm/arm-tables.opt: Regenerate.
6222 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6224 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
6225 for vector types other than V16QImode.
6226 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
6227 define_expand, and call altivec_expand_vec_perm_le when producing
6228 code with little endian element order.
6229 (*altivec_vperm_<mode>_internal): New insn having previous
6230 behavior of altivec_vperm_<mode>.
6231 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
6232 altivec_expand_vec_perm_le when producing code with little endian
6234 (*altivec_vperm_<mode>_uns_internal): New insn having previous
6235 behavior of altivec_vperm_<mode>_uns.
6237 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6239 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
6240 (altivec_vsumsws): Add handling for -maltivec=be with a little
6242 (altivec_vsumsws_direct): New.
6243 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
6244 gen_altivec_vsumsws.
6246 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6248 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
6249 vtable_pointer_value_to_binfo): New functions.
6250 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
6251 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
6253 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
6255 * config/nios2/nios2.md (load_got_register): Initialize GOT
6256 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
6257 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
6259 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6261 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
6262 preserverd by passthrough, do not propagate the type.
6264 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6266 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6267 (mips_atomic_assign_expand_fenv): New function.
6268 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6270 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6272 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6273 (__builtin_mips_set_fcsr): Likewise.
6274 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6275 MIPS_USI_FTYPE_VOID.
6276 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6277 (mips16_expand_set_fcsr): Likewise.
6278 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6279 (mips16_set_fcsr_stub): Likewise.
6280 (mips16_get_fcsr_one_only_stub): New class.
6281 (mips16_set_fcsr_one_only_stub): Likewise.
6282 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6283 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6284 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6285 (hard_float): New availability predicate.
6286 (mips_builtins): Add get_fcsr and set_fcsr.
6287 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6288 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6289 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6290 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6291 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6294 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6296 * config/mips/mips.c (mips_one_only_stub): New class.
6297 (mips_need_mips16_rdhwr_p): Replace with...
6298 (mips16_rdhwr_stub): ...this new variable.
6299 (mips16_stub_call_address): New function.
6300 (mips16_rdhwr_one_only_stub): New class.
6301 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6302 (mips_output_mips16_rdhwr): Delete.
6303 (mips_finish_stub): New function.
6304 (mips_code_end): Use it to handle rdhwr stubs.
6306 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6309 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6310 when calculating size of integer atomic types.
6312 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6314 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6316 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6318 PR tree-optimization/60003
6319 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6320 * profile.c (branch_prob): Use gimple_call_builtin_p
6321 to check for BUILT_IN_SETJMP_RECEIVER.
6322 * tree-inline.c (copy_bb): Call notice_special_calls.
6324 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6327 * lra-constraints.c (process_alt_operands): Update reload_sum only
6330 2014-01-31 Richard Henderson <rth@redhat.com>
6333 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6334 until after else_eh is processed.
6336 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6338 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6339 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6340 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6341 in smmintrin.h, remove them.
6342 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6343 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6344 * config/i386/i386.md (ROUND_SAE): Fix value.
6345 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6346 (const48_operand): New.
6347 * config/i386/subst.md (round), (round_expand): Use
6348 const_4_or_8_to_11_operand.
6349 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6351 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6353 * config/i386/constraints.md (Yk): Swap meaning with k.
6354 * config/i386/i386.md (movhi_internal): Change Yk to k.
6355 (movqi_internal): Ditto.
6356 (*k<logic><mode>): Ditto.
6359 (kandn<mode>): Ditto.
6360 (*<code>hi_1): Ditto.
6361 (*<code>qi_1): Ditto.
6362 (kxnor<mode>): Ditto.
6363 (kortestzhi): Ditto.
6364 (kortestchi): Ditto.
6366 (*one_cmplhi2_1): Ditto.
6367 (*one_cmplqi2_1): Ditto.
6368 * config/i386/sse.md (): Change k to Yk.
6369 (avx512f_load<mode>_mask): Ditto.
6370 (avx512f_blendm<mode>): Ditto.
6371 (avx512f_store<mode>_mask): Ditto.
6372 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6373 (avx512f_storedqu<mode>_mask): Ditto.
6374 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6376 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6377 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6378 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6379 (avx512f_maskcmp<mode>3): Ditto.
6380 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6381 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6382 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6383 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6384 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6385 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6386 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6387 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6388 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6389 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6390 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6391 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6392 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6393 (vec_extract_lo_<mode>_maskm): Ditto.
6394 (vec_extract_hi_<mode>_maskm): Ditto.
6395 (avx512f_vternlog<mode>_mask): Ditto.
6396 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6397 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6398 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6399 (avx512f_<code>v8div16qi2_mask): Ditto.
6400 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6401 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6402 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6403 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6404 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6405 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6406 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6407 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6408 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6409 (avx512cd_maskb_vec_dupv8di): Ditto.
6410 (avx512cd_maskw_vec_dupv16si): Ditto.
6411 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6412 (avx512f_vpermi2var<mode>3_mask): Ditto.
6413 (avx512f_vpermi2var<mode>3_mask): Ditto.
6414 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6415 (*avx512f_gathersi<mode>): Ditto.
6416 (*avx512f_gathersi<mode>_2): Ditto.
6417 (*avx512f_gatherdi<mode>): Ditto.
6418 (*avx512f_gatherdi<mode>_2): Ditto.
6419 (*avx512f_scattersi<mode>): Ditto.
6420 (*avx512f_scatterdi<mode>): Ditto.
6421 (avx512f_compress<mode>_mask): Ditto.
6422 (avx512f_compressstore<mode>_mask): Ditto.
6423 (avx512f_expand<mode>_mask): Ditto.
6424 * config/i386/subst.md (mask): Change k to Yk.
6425 (mask_scalar_merge): Ditto.
6428 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6430 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6432 2014-01-31 Richard Biener <rguenther@suse.de>
6435 * builtins.c (fold_builtin_memory_op): Make sure to not
6436 use a floating-point mode or a boolean or enumeral type for
6439 2014-01-30 DJ Delorie <dj@redhat.com>
6441 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6442 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6443 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6444 whenever main() has an epilogue.
6446 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6448 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6449 unused variable "field".
6450 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6451 (vsx_mergeh_<mode>): Likewise.
6452 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6453 (altivec_vmrghh): Likewise.
6454 (altivec_vmrghw): Likewise.
6455 (altivec_vmrglb): Likewise.
6456 (altivec_vmrglh): Likewise.
6457 (altivec_vmrglw): Likewise.
6458 (altivec_vspltb): Add missing uses.
6459 (altivec_vsplth): Likewise.
6460 (altivec_vspltw): Likewise.
6461 (altivec_vspltsf): Likewise.
6463 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6466 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6467 frame related instructions.
6469 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6471 PR rtl-optimization/59959
6472 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6473 any reload of register whose subreg is invalid.
6475 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6477 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6478 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6479 Add missing return type - void.
6481 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6483 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6484 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6485 remove element index adjustment for endian (now handled in vsx.md
6487 (altivec_expand_vec_perm_const): Use
6488 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6489 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6490 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6491 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6492 define_expand and a new define_insn *altivec_vspltb_internal;
6493 adjust for -maltivec=be on a little endian target.
6494 (altivec_vspltb_direct): New.
6495 (altivec_vsplth): Divide into a define_expand and a new
6496 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6497 little endian target.
6498 (altivec_vsplth_direct): New.
6499 (altivec_vspltw): Divide into a define_expand and a new
6500 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6501 little endian target.
6502 (altivec_vspltw_direct): New.
6503 (altivec_vspltsf): Divide into a define_expand and a new
6504 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6505 a little endian target.
6507 2014-01-30 Richard Biener <rguenther@suse.de>
6509 PR tree-optimization/59993
6510 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6511 can propagate form the earlier stmt and avoid the transform
6512 when the intermediate result is needed.
6514 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6516 * README.Portability: Fix typo.
6518 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6520 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6521 comparison_operator with ordered_comparison_operator.
6523 2014-01-30 Nick Clifton <nickc@redhat.com>
6525 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6526 Rename to mn10300_store_multiple_regs.
6527 * config/mn10300/mn10300.c: Likewise.
6528 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6529 store_multiple_regs.
6530 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6531 Call mn10300_store_multiple_regs.
6533 2014-01-30 Nick Clifton <nickc@redhat.com>
6534 DJ Delorie <dj@redhat.com>
6536 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6537 %fp 2 to keep registers after it properly word-aligned.
6538 (rl78_alloc_physical_registers_umul): Handle the case where both
6539 input operands are the same.
6541 2014-01-30 Richard Biener <rguenther@suse.de>
6543 PR tree-optimization/59903
6544 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6547 2014-01-30 Jason Merrill <jason@redhat.com>
6550 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6553 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6555 2014-01-30 Richard Biener <rguenther@suse.de>
6557 PR tree-optimization/59951
6558 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6560 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6563 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6564 SFmode to DFmode case.
6566 2014-01-29 DJ Delorie <dj@redhat.com>
6568 * config/msp430/msp430.opt (-minrt): New.
6569 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6571 (ENDFILE_SPEC): Likewise.
6573 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6575 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6576 (estimate_function_body_sizes): Use it.
6578 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6581 * dwarf2out.c (is_cxx_auto): New.
6582 (is_base_type): Use it.
6583 (gen_type_die_with_usage): Likewise.
6585 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6587 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6588 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6589 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6590 -maltivec=be with LE targets.
6591 (vsx_mergeh_<mode>): Likewise.
6592 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6593 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6594 (altivec_vmrghb): Replace with define_expand and new
6595 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6596 (altivec_vmrghb_direct): New define_insn.
6597 (altivec_vmrghh): Replace with define_expand and new
6598 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6599 (altivec_vmrghh_direct): New define_insn.
6600 (altivec_vmrghw): Replace with define_expand and new
6601 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6602 (altivec_vmrghw_direct): New define_insn.
6603 (*altivec_vmrghsf): Adjust for endianness.
6604 (altivec_vmrglb): Replace with define_expand and new
6605 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6606 (altivec_vmrglb_direct): New define_insn.
6607 (altivec_vmrglh): Replace with define_expand and new
6608 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6609 (altivec_vmrglh_direct): New define_insn.
6610 (altivec_vmrglw): Replace with define_expand and new
6611 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6612 (altivec_vmrglw_direct): New define_insn.
6613 (*altivec_vmrglsf): Adjust for endianness.
6614 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6615 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6616 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6617 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6618 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6619 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6620 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6621 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6623 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6625 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6626 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6629 2014-01-29 Richard Biener <rguenther@suse.de>
6631 PR tree-optimization/58742
6632 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6633 associate_pointerplus_align.
6634 (associate_pointerplus_diff): New function.
6635 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6636 and associate_pointerplus_diff.
6638 2014-01-29 Richard Biener <rguenther@suse.de>
6640 * lto-streamer.h (LTO_major_version): Bump to 3.
6641 (LTO_minor_version): Reset to 0.
6643 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6645 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6646 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6647 (arm_file_start): Generate correct asm header for armv7ve.
6648 * config/arm/bpabi.h: Add multilib support for armv7ve.
6649 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6650 and cortex-a15 to armv7ve.
6651 * config/arm/t-aprofile: Add multilib support for armv7ve.
6652 * doc/invoke.texi: Document -march=armv7ve.
6654 2014-01-29 Richard Biener <rguenther@suse.de>
6656 PR tree-optimization/58742
6657 * tree-ssa-forwprop.c (associate_plusminus): Return true
6658 if we changed sth, defer EH cleanup to ...
6659 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6660 (simplify_mult): New function.
6662 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6665 PR tree-optimization/59920
6666 * tree.c (build_common_builtin_nodes): Remove
6667 __builtin_setjmp_dispatcher initialization.
6668 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6669 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6670 instead of gsi_after_labels + manually skipping debug stmts.
6671 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6672 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6673 * tree-inline.c (copy_edges_for_bb): Remove
6674 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6675 argument. Ignore computed_goto_p stmts. Don't call
6676 make_abnormal_goto_edges. If a call might need abnormal edges
6677 for non-local gotos, see if it already has an edge to
6678 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6679 with true argument, don't do anything then, otherwise add
6680 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6681 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6683 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6684 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6685 (lower_stmt): Don't set data->calls_builtin_setjmp.
6686 (lower_builtin_setjmp): Adjust comment.
6687 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6688 * tree-cfg.c (found_computed_goto): Remove.
6689 (factor_computed_gotos): Remove.
6690 (make_goto_expr_edges): Return bool, true for computed gotos.
6691 Don't call make_abnormal_goto_edges.
6692 (build_gimple_cfg): Don't set found_computed_goto, don't call
6693 factor_computed_gotos.
6694 (computed_goto_p): No longer static.
6695 (make_blocks): Don't set found_computed_goto.
6696 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6697 (make_edges): If make_goto_expr_edges returns true, push bb
6698 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6699 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6700 vector. Record mapping between bbs and OpenMP regions if there
6701 are any, adjust make_gimple_omp_edges caller. Call
6702 handle_abnormal_edges.
6703 (make_abnormal_goto_edges): Remove.
6704 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6705 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6706 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6707 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6708 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6709 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6710 filling *region also set *region_idx to (*region)->entry->index.
6713 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6714 For REGs set ORIGINAL_REGNO.
6716 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6718 * doc/md.texi: Mention that a target shouldn't implement
6719 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6722 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6724 PR tree-optimization/59594
6725 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6726 a copy of the datarefs vector rather than the vector itself.
6728 2014-01-28 Jason Merrill <jason@redhat.com>
6731 * dwarf2out.c (auto_die): New static.
6732 (gen_type_die_with_usage): Handle C++1y 'auto'.
6733 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6736 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6739 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6740 (SPEC_X32): Likewise.
6741 (SPEC_64): Likewise.
6742 * config/i386/i386.c (ix86_option_override_internal): Turn off
6743 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6745 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6746 * config/i386/i386.h (TARGET_16BIT): New macro.
6747 (TARGET_16BIT_P): Likewise.
6748 * config/i386/i386.opt: Add m16.
6749 * doc/invoke.texi: Document -m16.
6751 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6753 PR preprocessor/59935
6754 * input.c (location_get_source_line): Bail out on when line number
6755 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6757 2014-01-28 Richard Biener <rguenther@suse.de>
6759 PR tree-optimization/58742
6760 * tree-ssa-forwprop.c (associate_plusminus): Handle
6761 pointer subtraction of the form (T)(P + A) - (T)P.
6763 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6765 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6768 2014-01-28 Richard Biener <rguenther@suse.de>
6771 2014-01-28 Richard Biener <rguenther@suse.de>
6773 PR rtl-optimization/45364
6774 PR rtl-optimization/59890
6775 * var-tracking.c (local_get_addr_clear_given_value): Handle
6776 already cleared slot.
6777 (val_reset): Handle not allocated local_get_addr_cache.
6778 (vt_find_locations): Use post-order on the inverted CFG.
6780 2014-01-28 Richard Biener <rguenther@suse.de>
6782 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6784 2014-01-28 Richard Biener <rguenther@suse.de>
6786 PR rtl-optimization/45364
6787 PR rtl-optimization/59890
6788 * var-tracking.c (local_get_addr_clear_given_value): Handle
6789 already cleared slot.
6790 (val_reset): Handle not allocated local_get_addr_cache.
6791 (vt_find_locations): Use post-order on the inverted CFG.
6793 2014-01-28 Alan Modra <amodra@gmail.com>
6795 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6796 * configure.ac <recursive call for build != host>: Define
6797 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6798 and LD_FOR_BUILD too.
6799 * configure: Regenerate.
6801 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6803 * config/i386/i386.c (get_builtin_code_for_version): Separate
6804 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6805 Broadwell from Haswell.
6807 2014-01-27 Steve Ellcey <sellcey@mips.com>
6809 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6810 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6811 * config/mips/mips.c (mips_option_override): Change setting
6813 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6814 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6815 Change from Mask to Var.
6817 2014-01-27 Jeff Law <law@redhat.com>
6819 * ipa-inline.c (inline_small_functions): Fix typo.
6821 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6823 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6824 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6825 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6826 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6827 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6828 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6829 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6830 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6831 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6832 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6833 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6834 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6835 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6836 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6837 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6838 (_mm512_storeu_epi64): Ditto.
6839 (_mm512_cmpge_epi32_mask): Ditto.
6840 (_mm512_cmpge_epu32_mask): Ditto.
6841 (_mm512_cmpge_epi64_mask): Ditto.
6842 (_mm512_cmpge_epu64_mask): Ditto.
6843 (_mm512_cmple_epi32_mask): Ditto.
6844 (_mm512_cmple_epu32_mask): Ditto.
6845 (_mm512_cmple_epi64_mask): Ditto.
6846 (_mm512_cmple_epu64_mask): Ditto.
6847 (_mm512_cmplt_epi32_mask): Ditto.
6848 (_mm512_cmplt_epu32_mask): Ditto.
6849 (_mm512_cmplt_epi64_mask): Ditto.
6850 (_mm512_cmplt_epu64_mask): Ditto.
6851 (_mm512_cmpneq_epi32_mask): Ditto.
6852 (_mm512_cmpneq_epu32_mask): Ditto.
6853 (_mm512_cmpneq_epi64_mask): Ditto.
6854 (_mm512_cmpneq_epu64_mask): Ditto.
6855 (_mm512_expand_pd): Ditto.
6856 (_mm512_expand_ps): Ditto.
6857 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6858 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6859 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6860 * config/i386/i386.c (ix86_builtins): Add
6861 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6862 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6863 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6864 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6865 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6866 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6867 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6868 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6869 IX86_BUILTIN_PMOVUSQW512_MEM.
6870 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6871 __builtin_ia32_pmovsqd512mem_mask,
6872 __builtin_ia32_pmovqd512mem_mask,
6873 __builtin_ia32_pmovusqw512mem_mask,
6874 __builtin_ia32_pmovsqw512mem_mask,
6875 __builtin_ia32_pmovqw512mem_mask,
6876 __builtin_ia32_pmovusdw512mem_mask,
6877 __builtin_ia32_pmovsdw512mem_mask,
6878 __builtin_ia32_pmovdw512mem_mask,
6879 __builtin_ia32_pmovqb512mem_mask,
6880 __builtin_ia32_pmovusqb512mem_mask,
6881 __builtin_ia32_pmovsqb512mem_mask,
6882 __builtin_ia32_pmovusdb512mem_mask,
6883 __builtin_ia32_pmovsdb512mem_mask,
6884 __builtin_ia32_pmovdb512mem_mask.
6885 (bdesc_args): Add __builtin_ia32_expanddf512,
6886 __builtin_ia32_expandsf512.
6887 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6888 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6889 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6890 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6891 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6892 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6893 (avx512f_<code>v8div16qi2_mask_store): This.
6894 (avx512f_expand<mode>): New.
6896 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6898 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6900 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6901 (_mm512_prefetch_i32scatter_pd): Ditto.
6902 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6903 (_mm512_prefetch_i64scatter_pd): Ditto.
6904 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6905 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6906 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6907 (_mm512_prefetch_i32scatter_ps): Ditto.
6908 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6909 (_mm512_prefetch_i64scatter_ps): Ditto.
6910 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6911 * config/i386/i386-builtin-types.def: Define
6912 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6913 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6914 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6915 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6916 IX86_BUILTIN_SCATTERPFQPD.
6917 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6918 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6919 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6920 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6921 __builtin_ia32_scatterpfqps.
6922 (ix86_expand_builtin): Expand new built-ins.
6923 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6924 fix memory access data type.
6925 (*avx512pf_gatherpf<mode>_mask): Ditto.
6926 (*avx512pf_gatherpf<mode>): Ditto.
6927 (avx512pf_scatterpf<mode>): Ditto.
6928 (*avx512pf_scatterpf<mode>_mask): Ditto.
6929 (*avx512pf_scatterpf<mode>): Ditto.
6930 (GATHER_SCATTER_SF_MEM_MODE): New.
6931 (avx512pf_gatherpf<mode>df): Ditto.
6932 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6933 (*avx512pf_scatterpf<mode>df): Ditto.
6935 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6938 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6939 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6942 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6944 * common/config/arm/arm-common.c
6945 (arm_rewrite_mcpu): Handle multiple names.
6947 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6949 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6951 * gimple-builder.h (create_gimple_tmp): Delete.
6953 2014-01-27 Christian Bruel <christian.bruel@st.com>
6955 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
6958 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
6960 * config/pa/pa.md (call): Generate indirect long calls to non-local
6961 functions when outputing 32-bit code.
6962 (call_value): Likewise except for special call to buggy powf function.
6964 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
6965 portable runtime and PIC indirect calls.
6966 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
6967 and PIC call sequences. Use ldo instead of blr to set return register
6968 in PIC call sequence.
6970 2014-01-25 Walter Lee <walt@tilera.com>
6972 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
6973 avoid clobbering a live register.
6975 2014-01-25 Walter Lee <walt@tilera.com>
6977 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
6978 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
6979 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
6980 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
6982 2014-01-25 Walter Lee <walt@tilera.com>
6984 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
6985 arguments on even registers.
6986 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
6988 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
6989 (BIGGEST_ALIGNMENT): Ditto.
6990 (BIGGEST_FIELD_ALIGNMENT): Ditto.
6992 2014-01-25 Walter Lee <walt@tilera.com>
6994 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
6995 insns before bundling.
6996 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
6998 2014-01-25 Walter Lee <walt@tilera.com>
7000 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
7001 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
7002 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
7004 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7006 * config/mips/constraints.md (kl): Delete.
7007 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
7008 define expands, using...
7009 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
7010 instructions for MIPS16.
7011 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
7012 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
7014 2014-01-25 Walter Lee <walt@tilera.com>
7016 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
7020 2014-01-25 Walter Lee <walt@tilera.com>
7022 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
7023 (TARGET_EXPAND_TO_RTL_HOOK): Define.
7025 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7027 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
7030 2014-01-25 Jakub Jelinek <jakub@redhat.com>
7032 * print-rtl.c (in_call_function_usage): New var.
7033 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
7034 EXPR_LIST mode as mode and not as reg note name.
7037 * cfgloopmanip.c (copy_loop_info): If
7038 loop->warned_aggressive_loop_optimizations, make sure
7039 the flag is set in target loop too.
7041 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
7043 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
7045 * builtins.def: Likewise.
7046 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
7047 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
7048 * ira.c (ira_setup_eliminable_regset): Likewise.
7049 * omp-low.c (gate_expand_omp): Likewise.
7050 (execute_lower_omp): Likewise.
7051 (diagnose_sb_0): Likewise.
7052 (gate_diagnose_omp_blocks): Likewise.
7053 (simd_clone_clauses_extract): Likewise.
7056 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7058 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
7059 correction for little endian...
7060 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
7063 2014-01-24 Jeff Law <law@redhat.com>
7065 PR tree-optimization/59919
7066 * tree-vrp.c (find_assert_locations_1): Do not register asserts
7067 for non-returning calls.
7069 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
7071 * common/config/aarch64/aarch64-common.c
7072 (aarch64_rewrite_mcpu): Handle multiple names.
7073 * config/aarch64/aarch64.h
7074 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
7076 2014-01-24 Dodji Seketeli <dodji@redhat.com>
7078 * input.c (add_file_to_cache_tab): Handle the case where fopen
7081 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
7084 * config/i386/i386.md (pushsf splitter): Get stack adjustment
7085 from push operand if code of push isn't PRE_DEC.
7087 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7090 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
7091 -mquad-memory-atomic. Update -mquad-memory documentation to say
7092 it is only used for non-atomic loads/stores.
7094 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
7095 -mquad-memory or -mquad-memory-atomic switches.
7097 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
7098 -mquad-memory-atomic to ISA 2.07 support.
7100 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
7101 to separate support of normal quad word memory operations (ldq, stq)
7102 from the atomic quad word memory operations.
7104 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
7105 support to separate non-atomic quad word operations from atomic
7106 quad word operations. Disable non-atomic quad word operations in
7107 little endian mode so that we don't have to swap words after the
7108 load and before the store.
7109 (quad_load_store_p): Add comment about atomic quad word support.
7110 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
7111 options printed with -mdebug=reg.
7113 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
7114 -mquad-memory-atomic as the test for whether we have quad word
7115 atomic instructions.
7116 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
7117 or -mp8-vector are used, allow byte/half-word atomic operations.
7119 * config/rs6000/sync.md (load_lockedti): Insure that the address
7120 is a proper indexed or indirect address for the lqarx instruction.
7121 On little endian systems, swap the hi/lo registers after the lqarx
7123 (load_lockedpti): Use indexed_or_indirect_operand predicate to
7124 insure the address is valid for the lqarx instruction.
7125 (store_conditionalti): Insure that the address is a proper indexed
7126 or indirect address for the stqcrx. instruction. On little endian
7127 systems, swap the hi/lo registers before doing the stqcrx.
7129 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
7130 insure the address is valid for the stqcrx. instruction.
7132 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
7133 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
7134 type of quad memory support is available.
7136 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
7139 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
7140 there is a danger of looping.
7142 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
7144 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
7145 force flag_ira_loop_pressure if set via command line.
7147 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7149 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
7150 (ashr_simd): New builtin handling DI mode.
7151 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
7152 (aarch64_sshr_simddi): New match pattern.
7153 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
7154 (vshrd_n_s64): Likewise.
7155 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
7157 2014-01-23 Nick Clifton <nickc@redhat.com>
7159 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
7160 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
7161 favour of mcu specific scripts.
7162 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
7165 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7166 Alex Velenko <Alex.Velenko@arm.com>
7168 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
7169 (vaddv_s16): Likewise.
7170 (vaddv_s32): Likewise.
7171 (vaddv_u8): Likewise.
7172 (vaddv_u16): Likewise.
7173 (vaddv_u32): Likewise.
7174 (vaddvq_s8): Likewise.
7175 (vaddvq_s16): Likewise.
7176 (vaddvq_s32): Likewise.
7177 (vaddvq_s64): Likewise.
7178 (vaddvq_u8): Likewise.
7179 (vaddvq_u16): Likewise.
7180 (vaddvq_u32): Likewise.
7181 (vaddvq_u64): Likewise.
7182 (vaddv_f32): Likewise.
7183 (vaddvq_f32): Likewise.
7184 (vaddvq_f64): Likewise.
7185 (vmaxv_f32): Likewise.
7186 (vmaxv_s8): Likewise.
7187 (vmaxv_s16): Likewise.
7188 (vmaxv_s32): Likewise.
7189 (vmaxv_u8): Likewise.
7190 (vmaxv_u16): Likewise.
7191 (vmaxv_u32): Likewise.
7192 (vmaxvq_f32): Likewise.
7193 (vmaxvq_f64): Likewise.
7194 (vmaxvq_s8): Likewise.
7195 (vmaxvq_s16): Likewise.
7196 (vmaxvq_s32): Likewise.
7197 (vmaxvq_u8): Likewise.
7198 (vmaxvq_u16): Likewise.
7199 (vmaxvq_u32): Likewise.
7200 (vmaxnmv_f32): Likewise.
7201 (vmaxnmvq_f32): Likewise.
7202 (vmaxnmvq_f64): Likewise.
7203 (vminv_f32): Likewise.
7204 (vminv_s8): Likewise.
7205 (vminv_s16): Likewise.
7206 (vminv_s32): Likewise.
7207 (vminv_u8): Likewise.
7208 (vminv_u16): Likewise.
7209 (vminv_u32): Likewise.
7210 (vminvq_f32): Likewise.
7211 (vminvq_f64): Likewise.
7212 (vminvq_s8): Likewise.
7213 (vminvq_s16): Likewise.
7214 (vminvq_s32): Likewise.
7215 (vminvq_u8): Likewise.
7216 (vminvq_u16): Likewise.
7217 (vminvq_u32): Likewise.
7218 (vminnmv_f32): Likewise.
7219 (vminnmvq_f32): Likewise.
7220 (vminnmvq_f64): Likewise.
7222 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7224 * config/aarch64/aarch64-simd.md
7225 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
7226 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
7227 (*aarch64_mul3_elt<mode>): Likewise.
7228 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
7229 (*aarch64_mul3_elt_to_64v2df): Likewise.
7230 (*aarch64_mla_elt<mode>): Likewise.
7231 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7232 (*aarch64_mls_elt<mode>): Likewise.
7233 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7234 (*aarch64_fma4_elt<mode>): Likewise.
7235 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7236 (*aarch64_fma4_elt_to_64v2df): Likewise.
7237 (*aarch64_fnma4_elt<mode>): Likewise.
7238 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7239 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7240 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7241 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7242 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7243 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7244 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7245 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7246 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7248 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
7250 * config/aarch64/aarch64-simd.md
7251 (aarch64_be_checked_get_lane<mode>): New define_expand.
7252 * config/aarch64/aarch64-simd-builtins.def
7253 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
7254 New builtin definition.
7255 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
7256 Use new safe be builtin.
7258 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7260 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
7262 (aarch64_be_st1<mode>): Likewise.
7263 (aarch_ld1<VALL:mode>): Define_expand modified.
7264 (aarch_st1<VALL:mode>): Likewise.
7265 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7266 (UNSPEC_ST1): Likewise.
7268 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
7270 * config/microblaze/microblaze.md: Add trap insn and attribute
7272 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7274 PR preprocessor/58580
7275 * input.h (location_get_source_line): Take an additional line_size
7277 (void diagnostics_file_cache_fini): Declare new function.
7278 * input.c (struct fcache): New type.
7279 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7280 New static constants.
7281 (diagnostic_file_cache_init, total_lines_num)
7282 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7283 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7284 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7285 (get_next_line, read_next_line, goto_next_line, read_line_num):
7286 New static function definitions.
7287 (diagnostic_file_cache_fini): New function.
7288 (location_get_source_line): Take an additional output line_len
7289 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7291 * diagnostic.c (diagnostic_finish): Call
7292 diagnostic_file_cache_fini.
7293 (adjust_line): Take an additional input parameter for the length
7294 of the line, rather than calculating it with strlen.
7295 (diagnostic_show_locus): Adjust the use of
7296 location_get_source_line and adjust_line with respect to their new
7297 signature. While displaying a line now, do not stop at the first
7298 null byte. Rather, display the zero byte as a space and keep
7299 going until we reach the size of the line.
7300 * Makefile.in: Add vec.o to OBJS-libcommon
7302 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7303 Ilya Tocar <ilya.tocar@intel.com>
7305 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7306 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7307 (__builtin_ia32_kmov16): Ditto.
7308 * config/i386/i386.md (UNSPEC_KMOV): New.
7311 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7313 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7314 (_mm512_storeu_si512): Ditto.
7316 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7319 * rtl.h (get_referenced_operands): Declare.
7320 * recog.c (get_referenced_operands): New function.
7321 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7322 operands have been referenced when recording LO_SUM references.
7324 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7326 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7328 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7330 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7331 Enable for generic and recent AMD targets.
7333 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7335 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7336 ARG_SIZE note when adjustment was eliminated.
7338 2014-01-22 Jeff Law <law@redhat.com>
7340 PR tree-optimization/59597
7341 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7342 in file. Accept new argument REGISTERING and use it to modify
7343 dump output appropriately.
7344 (register_jump_thread): Corresponding changes.
7345 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7346 thread paths involving joiner blocks. Add code to dump cancelled
7347 jump threading paths.
7349 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7351 PR rtl-optimization/59477
7352 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7353 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7355 2014-01-22 Tom Tromey <tromey@redhat.com>
7357 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7359 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7361 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7363 PR rtl-optimization/59896
7364 * lra-constraints.c (process_alt_operands): Check unused note for
7365 matched operands of insn with no output reloads.
7367 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7369 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7370 (mips_move_from_gpr_cost): Likewise.
7372 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7374 PR rtl-optimization/59858
7375 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7376 ira_class_hard_regs_num.
7377 (process_alt_operands): Increase reject for dying matched operand.
7379 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7382 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7383 smaller than size, perform several stores or loads and stores
7384 at dst + count - size to store or copy all of size bytes, rather
7385 than just last modesize bytes.
7387 2014-01-20 DJ Delorie <dj@redhat.com>
7389 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7390 that CLOBBERs are REGs before propogating their values.
7392 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7395 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7396 (cgraph_inline_failed_type): New function.
7397 * cgraph.h (DEFCIFCODE): Add type.
7398 (cgraph_inline_failed_type_t): New enum.
7399 (cgraph_inline_failed_type): New prototype.
7400 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7401 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7402 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7403 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7404 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7405 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7406 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7407 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7408 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7409 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7410 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7411 OPTIMIZATION_MISMATCH.
7412 * tree-inline.c (expand_call_inline): Emit errors during
7413 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7415 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7418 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7419 mode attribute in insn output.
7421 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7423 * output.h (output_constant): Delete.
7424 * varasm.c (output_constant): Make private.
7426 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7428 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7430 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7433 * tree.h (fold_builtin_strcat): New prototype.
7434 * builtins.c (fold_builtin_strcat): No longer static. Add len
7435 argument, if non-NULL, don't call c_strlen. Optimize
7436 directly into __builtin_memcpy instead of __builtin_strcpy.
7437 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7438 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7440 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7442 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7443 for SImode_address_operand operands, having only a REG argument.
7445 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7447 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7448 loader name using mbig-endian.
7449 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7451 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7453 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7457 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7459 * config/aarch64/aarch64-protos.h
7460 (aarch64_cannot_change_mode_class_ptr): Declare.
7461 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7462 aarch64_cannot_change_mode_class_ptr): New.
7463 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7464 backend hook aarch64_cannot_change_mode_class.
7466 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7468 * common/config/aarch64/aarch64-common.c
7469 (aarch64_handle_option): Don't handle any option order logic here.
7470 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7471 selected_cpu, warn on architecture version mismatch.
7472 (aarch64_override_options): Fix parsing order for option strings.
7474 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7475 Iain Sandoe <iain@codesourcery.com>
7478 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7479 warning. Amend comment to reflect current functionality.
7481 2014-01-20 Richard Biener <rguenther@suse.de>
7484 * builtins.c (fold_builtin_strcat): Remove case better handled
7485 by tree-ssa-strlen.c.
7487 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7489 * config/aarch64/aarch64.opt
7490 (mcpu, march, mtune): Make case-insensitive.
7492 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7495 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7496 if operands[1] is a REG or ZERO_EXTEND of a REG.
7498 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7500 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7502 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7504 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7505 long non-pic millicode calls.
7507 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7509 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7511 2014-01-19 Kito Cheng <kito@0xlab.org>
7513 * builtins.c (expand_movstr): Check movstr expand done or fail.
7515 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7516 H.J. Lu <hongjiu.lu@intel.com>
7519 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7520 to DImode for zero-extended addresses.
7522 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7524 PR rtl-optimization/57763
7525 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7526 on the new indirect jump_insn and increment LABEL_NUSES (label).
7528 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7532 * config.gcc (x86_archs): New variable.
7533 (x86_64_archs): Likewise.
7534 (x86_cpus): Likewise.
7535 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7536 --with-arch/--with-cpu= options.
7537 Support --with-arch=/--with-cpu={nehalem,westmere,
7538 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7540 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7542 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7543 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7545 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7547 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7549 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7552 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7553 clear cpp_get_options (parse_in)->warn_unused_macros for
7554 ix86_target_macros_internal with cpp_define.
7556 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7558 * jump.c (delete_related_insns): Keep (use (insn))s.
7559 * reorg.c (redundant_insn): Check for barriers too.
7561 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7563 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7565 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7567 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7568 call to $$dyncall when TARGET_LONG_CALLS is true.
7570 2014-01-17 Jeff Law <law@redhat.com>
7572 * ree.c (combine_set_extension): Temporarily disable test for
7573 changing number of hard registers.
7575 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7578 * ipa-inline-analysis.c (inline_free_summary):
7579 Do not free summary of aliases.
7581 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7584 * gimplify.c (gimplify_expr): Use create_tmp_var
7585 instead of create_tmp_var_raw. If cond doesn't have
7586 integral type, don't add the IFN_ANNOTATE builtin at all.
7588 2014-01-17 Martin Jambor <mjambor@suse.cz>
7591 * ipa-cp.c (prev_edge_clone): New variable.
7592 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7593 Also resize prev_edge_clone vector.
7594 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7595 (ipcp_edge_removal_hook): New function.
7596 (ipcp_driver): Register ipcp_edge_removal_hook.
7598 2014-01-17 Andrew Pinski <apinski@cavium.com>
7599 Steve Ellcey <sellcey@mips.com>
7602 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7605 2014-01-17 Jeff Law <law@redhat.com>
7608 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7609 so that pass_ccp runs first.
7611 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7613 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7614 (ix86_adjust_cost): Use !TARGET_XXX.
7615 (do_reorder_for_imul): Likewise.
7616 (swap_top_of_ready_list): Likewise.
7617 (ix86_sched_reorder): Likewise.
7619 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7621 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7622 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7623 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7624 (intel_memset): New. Duplicate slm_memset.
7625 (intel_cost): New. Duplicate slm_cost.
7626 (m_INTEL): New macro.
7627 (processor_target_table): Add "intel".
7628 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7629 with PROCESSOR_INTEL for "intel".
7630 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7631 PROCESSOR_SILVERMONT.
7632 (ix86_issue_rate): Likewise.
7633 (ix86_adjust_cost): Likewise.
7634 (ia32_multipass_dfa_lookahead): Likewise.
7635 (swap_top_of_ready_list): Likewise.
7636 (ix86_sched_reorder): Likewise.
7637 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7638 instead of TARGET_OPT_AGU.
7639 * config/i386/i386.h (TARGET_INTEL): New.
7640 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7641 (processor_type): Add PROCESSOR_INTEL.
7642 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7643 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7645 2014-01-17 Marek Polacek <polacek@redhat.com>
7648 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7651 2014-01-17 Richard Biener <rguenther@suse.de>
7653 PR tree-optimization/46590
7654 * opts.c (default_options_table): Add entries for
7655 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7656 all enabled at -O1 but not for -Og.
7657 * common.opt (fbranch-count-reg): Remove Init(1).
7658 (fmove-loop-invariants): Likewise.
7659 (ftree-pta): Likewise.
7661 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7663 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7664 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7665 decls to at least the GCC 4.8 used alignments.
7668 * tree-nested.c (convert_nonlocal_reference_stmt,
7669 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7670 of GIMPLE_BIND stmts, adjust associated decls.
7672 2014-01-17 Richard Biener <rguenther@suse.de>
7674 PR tree-optimization/46590
7675 * vec.h (vec<>::bseach): New member function implementing
7676 binary search according to C89 bsearch.
7677 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7678 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7679 bitmap pointer again. Make accesses_in_loop a flat array.
7680 (mem_ref_obstack): New global.
7681 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7682 (mark_ref_stored): Likewise.
7683 (ref_indep_loop_p_2): Likewise.
7684 (set_ref_stored_in_loop): New helper function.
7685 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7686 (memref_free): Adjust.
7687 (record_mem_ref_loc): Simplify.
7688 (gather_mem_refs_stmt): Adjust.
7689 (sort_locs_in_loop_postorder_cmp): New function.
7690 (analyze_memory_references): Sort accesses_in_loop after
7691 loop postorder number.
7692 (find_ref_loc_in_loop_cmp): New function.
7693 (for_all_locs_in_loop): Find relevant cluster of locs in
7694 accesses_in_loop and iterate without recursion.
7695 (execute_sm): Avoid uninit warning.
7696 (struct ref_always_accessed): Simplify.
7697 (ref_always_accessed::operator ()): Likewise.
7698 (ref_always_accessed_p): Likewise.
7699 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7700 loop postorder numbers here.
7701 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7704 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7707 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7708 on decls for which assemble_alias has been called.
7710 2014-01-17 Nick Clifton <nickc@redhat.com>
7712 * config/msp430/msp430.opt: (mcpu): New option.
7713 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7714 (msp430_option_override): Parse target_cpu. If the MCU name
7715 matches a generic string, clear target_mcu.
7716 (msp430_attr): Allow numeric interrupt values up to 63.
7717 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7718 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7720 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7722 * config/msp430/msp430.md (popm): Use %J rather than %I.
7723 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7724 (addhi_cy_i): Use immediate_operand for operand 2.
7725 * doc/invoke.texi: Document -mcpu option.
7727 2014-01-17 Richard Biener <rguenther@suse.de>
7729 PR rtl-optimization/38518
7730 * df.h (df_analyze_loop): Declare.
7731 * df-core.c: Include cfgloop.h.
7732 (df_analyze_1): Split out main part of df_analyze.
7733 (df_analyze): Adjust.
7734 (loop_inverted_post_order_compute): New function.
7735 (loop_post_order_compute): Likewise.
7736 (df_analyze_loop): New function avoiding whole-function
7738 * loop-invariant.c (find_defs): Use df_analyze_loop.
7739 (find_invariants): Adjust.
7740 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7742 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7744 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7745 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7747 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7749 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7750 traversal when removing references.
7752 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7755 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7757 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7760 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7761 pushing a reload for an autoinc when we had previously reloaded an
7762 inner part of the address.
7764 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7766 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7768 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7769 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7770 when not giving up or versioning for alias only because of
7772 (vect_analyze_data_ref_dependences): Set to true.
7773 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7775 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7776 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7780 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7783 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7784 operand 0 predicate for gathers, use a new pseudo as subtarget.
7786 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7789 * lra-constraints.c (process_alt_operands): Add printing debug info.
7790 Check absence of input/output reloads for matched operands too.
7792 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7794 PR rtl-optimization/59835
7795 * ira.c (ira_init_register_move_cost): Increase cost for
7798 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7800 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7802 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7805 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7806 non-register objects. Use gen_(high/low)part more consistently.
7809 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7812 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7813 endian support, remove tests for WORDS_BIG_ENDIAN.
7814 (p8_mfvsrd_3_<mode>): Likewise.
7815 (reload_gpr_from_vsx<mode>): Likewise.
7816 (reload_gpr_from_vsxsf): Likewise.
7817 (p8_mfvsrd_4_disf): Likewise.
7819 2014-01-16 Richard Biener <rguenther@suse.de>
7821 PR rtl-optimization/46590
7822 * lcm.c (compute_antinout_edge): Use postorder iteration.
7823 (compute_laterin): Use inverted postorder iteration.
7825 2014-01-16 Nick Clifton <nickc@redhat.com>
7828 * varasm.c (output_constant): Return the number of bytes actually
7830 (output_constructor_array_range): Update the field size with the
7831 number of bytes emitted by output_constant.
7832 (output_constructor_regular_field): Likewise. Also do not
7833 complain if the total number of bytes emitted is now greater
7834 than the expected fieldpos.
7835 * output.h (output_constant): Update prototype and descriptive comment.
7837 2014-01-16 Marek Polacek <polacek@redhat.com>
7840 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7841 it is error_mark_node.
7843 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7845 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7846 VALID_AVX256_REG_OR_OI_MODE.
7848 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7850 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7851 current procedure should be profiled.
7853 2014-01-15 Andrew Pinski <apinski@cavium.com>
7855 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7856 of moving from/to the STACK_REG register class.
7858 2014-01-15 Richard Henderson <rth@redhat.com>
7861 * reginfo.c (global_regs_decl): Globalize.
7862 * rtl.h (global_regs_decl): Declare.
7863 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7864 reserved via global_regs.
7866 2014-01-15 Teresa Johnson <tejohnson@google.com>
7868 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7870 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7872 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7873 and vmulosh rather than call gen_vec_widen_smult_*.
7874 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7875 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7876 (vec_widen_smult_even_v16qi): Likewise.
7877 (vec_widen_umult_even_v8hi): Likewise.
7878 (vec_widen_smult_even_v8hi): Likewise.
7879 (vec_widen_umult_odd_v16qi): Likewise.
7880 (vec_widen_smult_odd_v16qi): Likewise.
7881 (vec_widen_umult_odd_v8hi): Likewise.
7882 (vec_widen_smult_odd_v8hi): Likewise.
7883 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7884 vmuloub rather than call gen_vec_widen_umult_*.
7885 (vec_widen_umult_lo_v16qi): Likewise.
7886 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7887 vmulosb rather than call gen_vec_widen_smult_*.
7888 (vec_widen_smult_lo_v16qi): Likewise.
7889 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7890 rather than call gen_vec_widen_umult_*.
7891 (vec_widen_umult_lo_v8hi): Likewise.
7892 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7893 rather than call gen_vec_widen_smult_*.
7894 (vec_widen_smult_lo_v8hi): Likewise.
7896 2014-01-15 Jeff Law <law@redhat.com>
7898 PR tree-optimization/59747
7899 * ree.c (find_and_remove_re): Properly handle case where a second
7900 eliminated extension requires widening a copy created for elimination
7901 of a prior extension.
7902 (combine_set_extension): Ensure that the number of hard regs needed
7903 for a destination register does not change when we widen it.
7905 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7907 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7908 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7909 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7910 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7911 (avr-*-rtems*): Likewise.
7912 (bfin*-rtems*): Likewise.
7913 (moxie-*-rtems*): Likewise.
7914 (h8300-*-rtems*): Likewise.
7915 (i[34567]86-*-rtems*): Likewise.
7916 (lm32-*-rtems*): Likewise.
7917 (m32r-*-rtems*): Likewise.
7918 (m68k-*-rtems*): Likewise.
7919 (microblaze*-*-rtems*): Likewise.
7920 (mips*-*-rtems*): Likewise.
7921 (powerpc-*-rtems*): Likewise.
7922 (sh-*-rtems*): Likewise.
7923 (sparc-*-rtems*): Likewise.
7924 (sparc64-*-rtems*): Likewise.
7925 (v850-*-rtems*): Likewise.
7926 (m32c-*-rtems*): Likewise.
7928 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7930 PR rtl-optimization/59511
7931 * ira.c (ira_init_register_move_cost): Use memory costs for some
7932 cases of register move cost calculations.
7933 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7934 instead of BB frequency.
7935 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7936 * lra-assigns.c (find_hard_regno_for): Ditto.
7938 2014-01-15 Richard Biener <rguenther@suse.de>
7940 PR tree-optimization/59822
7941 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7942 (vectorizable_load): Use it to hoist defs of uses of invariant
7943 loads out of the loop.
7945 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7946 Kugan Vivekanandarajah <kuganv@linaro.org>
7949 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7952 2014-01-15 Richard Biener <rguenther@suse.de>
7954 PR rtl-optimization/59802
7955 * lcm.c (compute_available): Use inverted postorder to seed
7956 the initial worklist.
7958 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7961 * config/s390/s390.c (s390_preferred_reload_class): Don't return
7962 ADDR_REGS for invalid symrefs in non-PIC code.
7964 2014-01-15 Jakub Jelinek <jakub@redhat.com>
7967 * builtins.c (determine_block_size): Initialize *probable_max_size
7968 even if len_rtx is CONST_INT.
7970 2014-01-14 Andrew Pinski <apinski@cavium.com>
7972 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
7973 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
7974 (cortexa53_tunings): Likewise.
7975 (aarch64_sched_issue_rate): New function.
7976 (TARGET_SCHED_ISSUE_RATE): Define.
7978 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7980 * ira-costs.c (find_costs_and_classes): Add missed
7981 ira_init_register_move_cost_if_necessary.
7983 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7986 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
7988 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
7991 * config/i386/i386.c (type_natural_mode): Add a bool parameter
7992 to indicate if type is used for function return value. Warn ABI
7993 change if the vector mode isn't available for function return value.
7994 (ix86_function_arg_advance): Pass false to type_natural_mode.
7995 (ix86_function_arg): Likewise.
7996 (ix86_gimplify_va_arg): Likewise.
7997 (function_arg_32): Don't warn ABI change.
7998 (ix86_function_value): Pass true to type_natural_mode.
7999 (ix86_return_in_memory): Likewise.
8000 (ix86_struct_value_rtx): Removed.
8001 (TARGET_STRUCT_VALUE_RTX): Likewise.
8003 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8005 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
8006 converting a conditional jump into a conditional return.
8008 2014-01-14 Richard Biener <rguenther@suse.de>
8010 PR tree-optimization/58921
8011 PR tree-optimization/59006
8012 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
8013 hoisting invariant stmts.
8014 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
8015 invariant loads on the preheader edge if possible.
8017 2014-01-14 Joey Ye <joey.ye@arm.com>
8019 * doc/plugin.texi (Building GCC plugins): Update to C++.
8021 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
8023 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
8024 (_mm_rcp28_round_ss): Ditto.
8025 (_mm_rsqrt28_round_sd): Ditto.
8026 (_mm_rsqrt28_round_ss): Ditto.
8027 (_mm_rcp28_sd): Ditto.
8028 (_mm_rcp28_ss): Ditto.
8029 (_mm_rsqrt28_sd): Ditto.
8030 (_mm_rsqrt28_ss): Ditto.
8031 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
8032 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
8033 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
8034 (IX86_BUILTIN_RCP28SD): Ditto.
8035 (IX86_BUILTIN_RCP28SS): Ditto.
8036 (IX86_BUILTIN_RSQRT28SD): Ditto.
8037 (IX86_BUILTIN_RSQRT28SS): Ditto.
8038 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
8039 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
8040 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
8041 (ix86_expand_special_args_builtin): Expand new FTYPE.
8042 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
8043 (srcp14<mode>): Make insn unary.
8044 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
8045 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
8046 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
8047 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
8048 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
8049 Fix rounding: make it SAE only.
8050 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
8052 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
8054 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
8055 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
8056 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
8057 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
8058 (round_saeonly_mask_scalar_operand4): Ditto.
8059 (round_saeonly_mask_scalar_op3): Ditto.
8060 (round_saeonly_mask_scalar_op4): Ditto.
8062 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8064 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
8065 Implement -maltivec=be for vec_insert and vec_extract.
8067 2014-01-10 DJ Delorie <dj@redhat.com>
8069 * config/msp430/msp430.md (call_internal): Don't allow memory
8070 references with SP as the base register.
8071 (call_value_internal): Likewise.
8072 * config/msp430/constraints.md (Yc): New. For memory references
8073 that don't use SP as a base register.
8075 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
8076 "an integer without a # prefix"
8077 * config/msp430/msp430.md (epilogue_helper): Use it.
8079 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8082 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
8083 AVX512F gather builtins.
8084 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
8085 on gather decls with INTEGER_TYPE masktype.
8086 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
8087 directly into the builtin rather than hoisting it before loop.
8089 PR tree-optimization/59387
8090 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
8091 (scev_const_prop): If folded_casts and type has undefined overflow,
8092 use force_gimple_operand instead of force_gimple_operand_gsi and
8093 for each added stmt if it is assign with
8094 arith_code_with_undefined_signed_overflow, call
8095 rewrite_to_defined_overflow.
8096 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
8097 gimple-fold.h instead.
8098 (arith_code_with_undefined_signed_overflow,
8099 rewrite_to_defined_overflow): Moved to ...
8100 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
8101 rewrite_to_defined_overflow): ... here. No longer static.
8102 Include gimplify-me.h.
8103 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
8104 rewrite_to_defined_overflow): New prototypes.
8106 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8108 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
8110 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
8112 * builtins.c (get_object_alignment_2): Minor tweak.
8113 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
8115 2014-01-13 Christian Bruel <christian.bruel@st.com>
8117 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
8118 optimized non constant lengths.
8120 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8123 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
8124 load as __atomic_load_N if possible.
8126 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8128 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
8130 (rs6000_expand_builtin): Adjust call.
8132 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8135 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
8136 * config/rs6000/rs6000.c: Include target-globals.h.
8137 (rs6000_set_current_function): Instead of doing target_reinit
8138 unconditionally, use save_target_globals_default_opts and
8139 restore_target_globals.
8141 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
8143 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
8144 (rs6000_expand_builtin): Handle mffs and mtfsf.
8145 (rs6000_init_builtins): Define mffs and mtfsf.
8146 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
8147 (rs6000_mffs): New pattern.
8148 (rs6000_mtfsf): New pattern.
8150 2014-01-11 Bin Cheng <bin.cheng@arm.com>
8152 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
8153 Start narrowing with START. Apply candidate-use pair
8154 and check overall cost in narrowing.
8155 (iv_ca_prune): Pass new argument.
8157 2014-01-10 Jeff Law <law@redhat.com>
8160 * ree.c (combine_reaching_defs): Ensure the defining statement
8161 occurs before the extension when optimizing extensions with
8162 different source and destination hard registers.
8164 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8167 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
8168 vtables into the type inheritance graph.
8170 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8172 PR rtl-optimization/59754
8173 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
8174 modes in the REGNO != REGNO case.
8176 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8178 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
8180 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8182 PR tree-optimization/59745
8183 * tree-predcom.c (tree_predictive_commoning_loop): Call
8184 free_affine_expand_cache if giving up because components is NULL.
8186 * target-globals.c (save_target_globals): Allocate < 4KB structs using
8187 GC in payload of target_globals struct instead of allocating them on
8188 the heap and the larger structs separately using GC.
8189 * target-globals.h (struct target_globals): Make regs, hard_regs,
8190 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
8191 of GTY((skip)) and change type to void *.
8192 (reset_target_globals): Cast loads from those fields to corresponding
8195 2014-01-10 Steve Ellcey <sellcey@mips.com>
8198 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
8199 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
8200 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
8202 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
8205 * aarch64-modes.def (CC_Zmode): New flags mode.
8206 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
8207 represents an equality.
8208 (aarch64_get_condition_code): Handle CC_Zmode.
8209 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
8211 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8213 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
8214 extraction in good case.
8216 2014-01-10 Richard Biener <rguenther@suse.de>
8218 PR tree-optimization/59374
8219 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
8220 checking after SLP discovery. Mark stmts not participating
8221 in any SLP instance properly.
8223 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8225 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
8226 when handling a SET rtx.
8228 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8230 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
8231 (cortex-a57): Likewise.
8232 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
8234 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8236 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
8237 non-iwmmxt builtins.
8239 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8243 * ipa-devirt.c record_target_from_binfo): Take as argument
8244 stack of binfos and lookup matching one for virtual inheritance.
8245 (possible_polymorphic_call_targets_1): Update.
8247 2014-01-10 Huacai Chen <chenhc@lemote.com>
8249 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
8250 kernel strings for Loongson-2E/2F/3A.
8252 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8255 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
8256 is_gimple_call before calling gimple_call_internal_p.
8258 2014-01-09 Steve Ellcey <sellcey@mips.com>
8260 * Makefile.in (TREE_FLOW_H): Remove.
8261 (TREE_SSA_H): Add file names from tree-flow.h.
8262 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
8263 * tree.h: Remove tree-flow.h reference.
8264 * hash-table.h: Remove tree-flow.h reference.
8265 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8266 reference with tree-ssa-loop.h.
8268 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8270 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8271 default element-order behavior for -maltivec.
8272 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8273 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8274 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8275 when targeting big endian, at least for now.
8276 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8278 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8281 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8282 var satisfies use_register_for_decl, just take into account type
8283 alignment, rather than decl alignment.
8285 PR tree-optimization/59622
8286 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8287 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8288 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8289 Don't devirtualize for inplace at all. For targets.length () == 1,
8290 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8292 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8294 * config/i386/i386.md (cpu): Remove the unused btver1.
8296 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8298 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8300 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8303 * tree-core.h (struct target_globals): New forward declaration.
8304 (struct tree_target_option): Add globals field.
8305 * tree.h (TREE_TARGET_GLOBALS): Define.
8306 (prepare_target_option_nodes_for_pch): New prototype.
8307 * target-globals.h (struct target_globals): Define even if
8309 * tree.c (prepare_target_option_node_for_pch,
8310 prepare_target_option_nodes_for_pch): New functions.
8311 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8312 * config/i386/i386.c: Include target-globals.h.
8313 (ix86_set_current_function): Instead of doing target_reinit
8314 unconditionally, use save_target_globals_default_opts and
8315 restore_target_globals.
8317 2014-01-09 Richard Biener <rguenther@suse.de>
8319 PR tree-optimization/59715
8320 * tree-cfg.h (split_critical_edges): Declare.
8321 * tree-cfg.c (split_critical_edges): Export.
8322 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8324 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8326 * cfgexpand.c (expand_stack_vars): Optionally disable
8327 asan stack protection.
8328 (expand_used_vars): Likewise.
8329 (partition_stack_vars): Likewise.
8330 * asan.c (asan_emit_stack_protection): Optionally disable
8331 after return stack usage.
8332 (instrument_derefs): Optionally disable memory access instrumentation.
8333 (instrument_builtin_call): Likewise.
8334 (instrument_strlen_call): Likewise.
8335 (asan_protect_global): Optionally disable global variables protection.
8336 * doc/invoke.texi: Added doc for new options.
8337 * params.def: Added new options.
8338 * params.h: Likewise.
8340 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8342 PR rtl-optimization/59724
8343 * ifcvt.c (cond_exec_process_if_block): Don't call
8344 flow_find_head_matching_sequence with 0 longest_match.
8345 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8346 non-active insns if !stop_after.
8347 (try_head_merge_bb): Revert 2014-01-07 changes.
8349 2014-01-08 Jeff Law <law@redhat.com>
8351 * ree.c (get_sub_rtx): New function, extracted from...
8352 (merge_def_and_ext): Here.
8353 (combine_reaching_defs): Use get_sub_rtx.
8355 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8357 * cgraph.h (varpool_variable_node): Do not choke on null node.
8359 2014-01-08 Catherine Moore <clm@codesourcery.com>
8361 * config/mips/mips.md (simple_return): Attempt to use JRC
8363 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8365 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8367 PR rtl-optimization/59137
8368 * reorg.c (steal_delay_list_from_target): Call update_block for
8370 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8372 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8374 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8375 two duplicate entries.
8377 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8380 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8382 * config/mips/mips.c (mips_truncated_op_cost): New function.
8383 (mips_rtx_costs): Adjust test for BADDU.
8384 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8386 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8388 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8389 (*baddu_si): ...this new pattern.
8391 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8394 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8396 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8399 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8401 (expand_expr, expand_normal): Adjust.
8402 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8403 inner_reference_p. Use inner_reference_p to expand inner references.
8404 (store_expr): Adjust.
8405 * cfgexpand.c (expand_call_stmt): Adjust.
8407 2014-01-08 Rong Xu <xur@google.com>
8409 * gcov-io.c (gcov_var): Move from gcov-io.h.
8410 (gcov_position): Ditto.
8411 (gcov_is_error): Ditto.
8412 (gcov_rewrite): Ditto.
8413 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8414 only part to libgcc/libgcov.h.
8416 2014-01-08 Marek Polacek <polacek@redhat.com>
8419 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8421 2014-01-08 Marek Polacek <polacek@redhat.com>
8424 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8426 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8428 PR rtl-optimization/59649
8429 * stor-layout.c (get_mode_bounds): For BImode return
8430 0 and STORE_FLAG_VALUE.
8432 2014-01-08 Richard Biener <rguenther@suse.de>
8435 * gimple.h (is_gimple_builtin_call): Remove.
8436 (gimple_builtin_call_types_compatible_p): New.
8437 (gimple_call_builtin_p): New overload.
8438 * gimple.c (is_gimple_builtin_call): Remove.
8439 (validate_call): Rename to ...
8440 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8442 (validate_type): New static function.
8443 (gimple_call_builtin_p): New overload and adjust.
8444 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8445 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8446 (gimple_fold_stmt_to_constant_1): Likewise.
8447 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8449 2014-01-08 Richard Biener <rguenther@suse.de>
8452 * gimplify.c (gimplify_expr): Gimplify register-register type
8453 VIEW_CONVERT_EXPRs to separate stmts.
8455 2014-01-07 Jeff Law <law@redhat.com>
8458 * ree.c (combine_set_extension): Handle case where source
8459 and destination registers in an extension insn are different.
8460 (combine_reaching_defs): Allow source and destination registers
8461 in extension to be different under limited circumstances.
8462 (add_removable_extension): Remove restriction that the
8463 source and destination registers in the extension are the same.
8464 (find_and_remove_re): Emit a copy from the extension's
8465 destination to its source after the defining insn if
8466 the source and destination registers are different.
8469 * ifcvt.c (merge_if_block): If we are merging a block with more than
8470 one successor with a block with no successors, remove any BARRIER
8471 after the second block.
8473 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8475 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8477 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8480 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8481 for 14-bit register offsets when INT14_OK_STRICT is false.
8483 2014-01-07 Roland Stigge <stigge@antcom.de>
8484 Michael Meissner <meissner@linux.vnet.ibm.com>
8487 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8488 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8490 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8492 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8495 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8497 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8498 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8499 rtx is const0_rtx or not.
8501 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8504 * target-globals.c (save_target_globals): Remove this_fn_optab
8506 * toplev.c: Include optabs.h.
8507 (target_reinit): Temporarily restore the global options if another
8508 set of options are in force.
8510 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8512 PR rtl-optimization/58668
8513 * cfgcleanup.c (flow_find_cross_jump): Don't count
8514 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8515 to determine what is counted.
8516 (flow_find_head_matching_sequence): Use active_insn_p to determine
8518 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8520 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8521 determine what is counted.
8523 PR tree-optimization/59643
8524 * tree-predcom.c (split_data_refs_to_components): If one dr is
8525 read and one write, determine_offset fails and the write isn't
8526 in the bad component, just put the read into the bad component.
8528 2014-01-07 Mike Stump <mikestump@comcast.net>
8529 Jakub Jelinek <jakub@redhat.com>
8532 * tree-core.h (struct tree_optimization_option): Change optabs
8533 type from unsigned char * to void *.
8534 * optabs.c (init_tree_optimization_optabs): Adjust
8535 TREE_OPTIMIZATION_OPTABS initialization.
8537 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8540 * config/i386/i386.h (struct machine_function): Add
8541 no_drap_save_restore field.
8542 * config/i386/i386.c (ix86_save_reg): Use
8543 !cfun->machine->no_drap_save_restore instead of
8544 crtl->stack_realign_needed.
8545 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8546 this function clears frame_pointer_needed. Set
8547 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8548 and DRAP reg is needed.
8550 2014-01-06 Marek Polacek <polacek@redhat.com>
8553 * doc/implement-c.texi: Mention that other integer types are
8554 permitted as bit-field types in strictly conforming mode.
8556 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8558 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8561 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8563 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8565 2014-01-06 Martin Jambor <mjambor@suse.cz>
8568 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8570 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8572 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8576 * var-tracking.c (add_stores): Preserve the value of the source even if
8577 we don't record the store.
8579 2014-01-06 Terry Guo <terry.guo@arm.com>
8581 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8583 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8586 * config/darwin.c (darwin_function_section): Adjust return values to
8587 correspond to optimisation changes made in r206070.
8589 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8591 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8592 from prefetch_block tune setting.
8593 (nocona_cost): Correct size of prefetch block to 64.
8595 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8597 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8598 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8599 used to save the static chain register in the computation of the offset
8600 from which the FP registers need to be restored.
8602 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8604 PR tree-optimization/59519
8605 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8606 ICE if get_current_def (current_new_name) is already non-NULL, as long
8607 as it is a phi result of some other phi in *new_exit_bb that has
8610 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8611 or vmovdqu* for misaligned_operand.
8612 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8613 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8614 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8615 aligned_mem for AVX512F masked aligned load and store builtins and for
8618 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8620 PR tree-optimization/59651
8621 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8622 Address range for negative step should be added by TYPE_SIZE_UNIT.
8624 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8626 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8627 overlapping registers also for registers other than the stack pointer.
8629 2014-01-03 Marek Polacek <polacek@redhat.com>
8632 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8635 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8638 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8641 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8642 (push splitter): Use <P:MODE_SIZE> instead of
8643 GET_MODE_SIZE (<P:MODE>mode).
8644 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8645 (mov -1, reg peephole2): Likewise.
8646 * config/i386/sse.md (*mov<mode>_internal,
8647 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8648 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8649 *<code><mode>3, *andnot<mode>3<mask_name>,
8650 <mask_codefor><code><mode>3<mask_name>): Likewise.
8651 * config/i386/subst.md (mask_mode512bit_condition,
8652 sd_mask_mode512bit_condition): Likewise.
8654 2014-01-02 Xinliang David Li <davidxl@google.com>
8656 PR tree-optimization/59303
8657 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8658 (dump_predicates): Better output format.
8659 (pred_equal_p): New function.
8660 (is_neq_relop_p): Ditto.
8661 (is_neq_zero_form_p): Ditto.
8662 (pred_expr_equal_p): Ditto.
8663 (pred_neg_p): Ditto.
8664 (simplify_pred): Ditto.
8665 (simplify_preds_2): Ditto.
8666 (simplify_preds_3): Ditto.
8667 (simplify_preds_4): Ditto.
8668 (simplify_preds): Ditto.
8670 (push_to_worklist): Ditto.
8671 (get_pred_info_from_cmp): Ditto.
8672 (is_degenerated_phi): Ditto.
8673 (normalize_one_pred_1): Ditto.
8674 (normalize_one_pred): Ditto.
8675 (normalize_one_pred_chain): Ditto.
8676 (normalize_preds): Ditto.
8677 (normalize_cond_1): Remove function.
8678 (normalize_cond): Ditto.
8679 (is_gcond_subset_of): Ditto.
8680 (is_subset_of_any): Ditto.
8681 (is_or_set_subset_of): Ditto.
8682 (is_and_set_subset_of): Ditto.
8683 (is_norm_cond_subset_of): Ditto.
8684 (pred_chain_length_cmp): Ditto.
8685 (convert_control_dep_chain_into_preds): Type change.
8686 (find_predicates): Ditto.
8687 (find_def_preds): Ditto.
8688 (destroy_predicates_vecs): Ditto.
8689 (find_matching_predicates_in_rest_chains): Ditto.
8690 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8691 (is_pred_expr_subset): Ditto.
8692 (is_pred_chain_subset_of): Ditto.
8693 (is_included_in): Ditto.
8694 (is_superset_of): Ditto.
8696 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8698 Update copyright years.
8700 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8702 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8703 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8704 config/arc/arc.md, config/arc/arc.opt,
8705 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8706 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8707 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8708 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8709 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8710 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8711 vtable-verify.c, vtable-verify.h: Use the standard form for the
8714 2014-01-02 Tobias Burnus <burnus@net-b.de>
8716 * gcc.c (process_command): Update copyright notice dates.
8717 * gcov-dump.c: Ditto.
8719 * doc/cpp.texi: Bump @copying's copyright year.
8720 * doc/cppinternals.texi: Ditto.
8721 * doc/gcc.texi: Ditto.
8722 * doc/gccint.texi: Ditto.
8723 * doc/gcov.texi: Ditto.
8724 * doc/install.texi: Ditto.
8725 * doc/invoke.texi: Ditto.
8727 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8729 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8731 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8733 * config/i386/sse.md (*mov<mode>_internal): Guard
8734 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8736 PR rtl-optimization/59647
8737 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8738 new_rtx into UNSIGNED_FLOAT rtxes.
8740 Copyright (C) 2014 Free Software Foundation, Inc.
8742 Copying and distribution of this file, with or without modification,
8743 are permitted in any medium without royalty provided the copyright
8744 notice and this notice are preserved.